WO2003067494A1 - Agents intelligents de negotiation ouverts a des demandes situes dans un systeme reparti - Google Patents

Agents intelligents de negotiation ouverts a des demandes situes dans un systeme reparti Download PDF

Info

Publication number
WO2003067494A1
WO2003067494A1 PCT/US2001/047023 US0147023W WO03067494A1 WO 2003067494 A1 WO2003067494 A1 WO 2003067494A1 US 0147023 W US0147023 W US 0147023W WO 03067494 A1 WO03067494 A1 WO 03067494A1
Authority
WO
WIPO (PCT)
Prior art keywords
buyer
seller
agent
intelligent
negotiation
Prior art date
Application number
PCT/US2001/047023
Other languages
English (en)
Inventor
Neal Solomon
Original Assignee
Neal Solomon
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Neal Solomon filed Critical Neal Solomon
Priority to AU2001297749A priority Critical patent/AU2001297749A1/en
Priority to GB0314327A priority patent/GB2390194A/en
Publication of WO2003067494A1 publication Critical patent/WO2003067494A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/08—Auctions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising

Definitions

  • the invention relates to an integrated automated commercial system and the apparatus and methods thereof.
  • novel databases, data search approaches, data mining, data analysis and data synthesis methods are used to provide a system for conducting disintermediated, point-to-point electronic commerce. Rivers of data are continuously automatically analyzed, and data is selected for inclusion in marketing promotion-driven showcase databases.
  • Buyer driven search queries initiate the commercial negotiation process by which multiple sellers simultaneously compete for orders in a multivariate way (beyond price alone). Prospective buyers and sellers may collaborate prior to an initial search petition.
  • the system uses intelligent mobile software agents to analyze, search, negotiate for and complete commercial transactions. To retain independence, the agents are endowed with complex adaptive artificial intelligence capabilities.
  • Intelligent Negotiation Agents INAs
  • Custom search and custom production capabilities are also implemented, thereby allowing increasingly efficient made-to-order services.
  • MAS multi-agent system
  • CCNs vertical industry cooperatives [cooperative communications networks]
  • AAs analytical agents
  • ITAs intelligent transaction agents
  • INAs INAs with ITAs
  • the cluster of inventions comprising the present system represents the consolidation of solutions to large economic systems integration problems.
  • Background of the Present Invention The emergence of the Internet has caused a shift in the methods of commercial activity towards automated purchasing, marketing, sales and distribution of products, services and bundles.
  • the automated aspects of electronic commerce allow a one-to-one relationship between seller and buyer as compared with the traditional mass production and sales approach.
  • most electronic commerce sales systems resemble simple catalog sales or intermediated exchange: Neither of these main approaches satisfies the ideal of automated commerce.
  • the present invention derives from the convergence of several technologies involving distributed computing systems and multi-agent systems.
  • GRID distributed database architecture
  • the new search technologies overcome these problems by exploiting distributed computing architectures and object-relational data structures.
  • EDI electronic data interchange
  • a technology that is emerging to succeed EDI involves a new programming language - XML - and business registry - UDDI. These new technologies allow a more robust communication between businesses because products, services and bundles are indexed and catalogued for direct access. Though more robust than EDI, XML/UDDI systems are merely passive information-based formats that link businesses, similar to the yellow-pages. Technologies to connect machines and people have been advanced by the advent of graphic user interface (GUI) technology applied to PCs through advanced operating systems. In addition to simpler GUIs, translation software has been used to bridge the gaps between different software applications. The development of a new generation of inter-agents that interface with human users and computer programs is a key evolution towards simpler, yet more powerful commercial transactions.
  • GUI graphic user interface
  • Multi-Agent Systems are not new in academic circles. The attempt to develop MASs in distributed computing environments has been active for over a decade. With the increased automation of business computing systems, MASs have reshaped the factory floor, securities trading systems and complex communications networks. With the advent of Al technologies, most prominently GP, GA, NN and FL, MASs have emerged as a reborn technology category for computer scientists. One of the most practical uses for MASs applies to negotiation systems. The development of the "contract net" (k-net) system for distributed problem solving acts as the pioneer idea for mediating distributed computer encounters, particularly for efficiently dividing limited computation resources in a network. Building on the k-net platform, prominent based market models for transaction negotiation include the
  • Organizing cooperating intelligent agents is a key challenge of computer science, because it involves calibrating rules that provide neutrality to the coordination (typically of buyers) in complex self-organizing systems.
  • automated systems that emulate manual transaction processes, most approaches merely represent evolutionary progress in the field.
  • E-commerce approaches seek to integrate post-sale systems with payment processing and CRM systems. This is necessary to complete and track transactions and to develop and enhance personalized customer relationships in a single centralized system. Even more advanced automated systems can more fully integrate complex marketing and financial processes into the transaction system.
  • the transaction system can be part of a unified system that includes data analysis and synthesis and negotiation processes. Summary of the Main Embodiments of the Invention
  • the present invention consists of two interdependent systems: (I) A network operating system for databases, database search, data analysis and synthesis, database inter-agents and data collaboration, and (II) A multi-agent system for negotiation and completion of transactions between parties.
  • the first system consists of Cooperative Communications Networks (CCNs) that are comprised of (primarily vertical) industry participants.
  • CCNs Cooperative Communications Networks
  • Participants use database showcases to stream data on products, services and bundles, continuously in real time.
  • Showcase databases use showcase inter-agents to automate the item selection process; such inter-agents access analyses of market trends and behavior to make item selections for inclusion into a showcase.
  • inter-agents access analyses of market trends and behavior to make item selections for inclusion into a showcase.
  • CCNs may be horizontal or customized: Such configurations can be buyer biased, such as a very large corporation automatically sourcing vendor orders.
  • Showcases are accessed by commercial search agents (CSAs). Because the showcases of each vertical industry are continuously updated, the search process is both fast and accurate.
  • the CSA uses information obtained from data mining processes to focus the search request.
  • the CSA acts as an initial commercial search query in most cases. Further negotiation processes follow the initial search after ranking search results according to buyer priorities.
  • CSAs can make requests based on numerous variables beyond price alone.
  • AAs Analytical agents
  • AAs are employed at the database system level for data mining, data analysis and data synthesis. AAs get continuous data inputs of general economic and market trends as well as company and product/service information. AAs have several functions, including making recommendations by using advanced collaborative filtering techniques. In addition, AAs synthesize information in the form of producing customized reports. Furthermore, AAs access services such as credit and accountability indices, finance and insurance opportunities, RMOs, promotions and computational resources. Such information breadth makes AAs an integral data computation resource for other agents in the system, most prominently ITAs and INAs.
  • AAs use Evolutionary Computation (EC) technologies in order to develop economic scenario forecasts.
  • EC Evolutionary Computation
  • genetic programming (GP) approaches are used, as well as genetic algorithms (GA) and neural network (NN) methods, that compare the constantly changing market conditions with customer preferences and provide adaptive real time analysis and customized advice.
  • GA genetic algorithms
  • NN neural network
  • CCNs cooperative communications networks
  • CSAs are free for basic services but can access AA services.
  • AAs have various levels of services that are accessible by users for supplemental fees.
  • the system uses codes to transfer information. These codes may be processed using languages such as the extensible mark-up language (XML) and registries (UDDI, RDF) as well as proprietary information exchange methods (SOAP). Some of the mobile program codes are written in the Java, Java 2, Java Beans, Jini, C++, C# and other languages. Inter-agents are used to perform functions between human and machine. For instance, showcase (or seller) inter-agents (S-IAs) automate the continuous updating of showcase databases. Buyer inter- agents (B-IAs) are also used to interface between users and their CSAs and INAs.
  • the multi-agent system (MAS) is the core system and process for the negotiation and completion of transactions.
  • the MAS consists of intelligent negotiation agents (INAs) and intelligent transaction agents (ITAs).
  • INAs have buyer (b-INA) or seller (s-INA) roles; similarly, ITAs have b-ITA and s-ITA roles.
  • CSA has initiated a search query for information (and promotions) to CCN showcase databases
  • at least two s-INAs respond with an initial ask price, as well as alternative prices for different product or service features, quality, quantity, delivery times, etc.
  • the buyer may request bidding information about bundles of products and services as well as individual items.
  • the multivariate negotiation process commences once a b-INA is launched to interact with the s-INAs.
  • the multilateral negotiation process occurs when the b-INA negotiates with multiple vendor s-INAs simultaneously. Such INA interactions may occur in parallel at buyer or seller locations. Multiple sellers are eventually limited to two sellers per session until ultimately one is selected by the buyer.
  • b-INAs initially negotiate simultaneously with at least two s-INAs, and because the INAs are mobile (and cannot be in two places simultaneously), in order to overcome latency lags b-INAs launch micro- agents that complete simultaneously interactive negotiations with multiple s-INAs
  • INAs use negotiation, auction and pricing strategy modules to establish, modify, evaluate and respond to bids. Further, specific approaches are used to conceal negotiation strategies, particularly time- based methods. Additionally, INAs can employ various "personalities" on a spectrum of attitudes in order to accelerate or decelerate the negotiations.
  • INAs are informed by AAs, which provide data analysis and synthesis functions, such as collaborative filtering-based recommendations, scenario forecasts and trend histories, that are crucial for effective negotiations.
  • INAs themselves employ artificial intelligence (Al) technologies. These autonomous agents use evolutionary computation methods in which computer programs learn and adapt to the changing commercial environment.
  • the main evolutionary computation approaches include
  • INAs artificial intelligence by INAs produces autonomous agents and self-organizing commercial systems.
  • the multi-agent system resembles a road system with various autonomous cars operating simultaneously.
  • the Al uses recognized rules for cars to interact, yet provides enough independence between each autonomous vehicle that all functions are not pre-destined.
  • Each car has its own endowments of power and efficiency as well as starting and destination points in space and time.
  • Each operates both within the limits of varying roadways and road conditions.
  • the overall system operates according to rules that allow an optimized flow of mobile activity. Yet, because agents have complex and changing priorities, they have varying associations. Taken together, the agents create a dynamic system that adapts as conditions and priorities change.
  • INAs operate in a computer system by sending program code and data between machines to fulfill a goal of completing a transaction.
  • the negotiation agents themselves move between machines.
  • Negotiations can occur between agents at specific locations, at multiple locations or between alternating locations.
  • INA mobility involves replicating program code, satisfying security protocols, pruning program code for increased mobility, retrieving layers of Al computation resources when needed, integrating essential database functions and accessing updated programming instructions from a home port. Mobility has numerous advantages for participants, including efficiency enhancements and operation in a system with communication constraints.
  • Cooperative INAs are comprised of groups of buyer INAs that band together in various ways in order to negotiate optimal deals.
  • C-INAs can be used for the aggregation of buyers for the acquisition of multi-item bundles that can be customized so that specific items in each customer package are individually tailored.
  • C-INAs allow the disintermediation of a wholesale layer in the distribution and production system by streamlining the sales process and by also providing discount buying power.
  • c-INAs are typically either buyer- or seller- biased.
  • B-C-INAs emphasize aggregation operations.
  • S-C-INAs are used as sellers must provisionally cooperate in order to calculate buyer values, particularly for the purpose of selling combination item bundles.
  • Pre-established multi-item bundles such as a pre-configured combination of computer hardware and software, can be treated as a single item for the purposes of this system.
  • open bundles consisting of multiple items require the selection of the buyer to assemble, and involve much more complex negotiations.
  • multi-item bundle bidding has emphasized the sale of multiple items from one seller to multiple buyers (such as an FCC spectrum auction).
  • the present system accommodates both a single buyer bidder with multiple sellers as well as multiple buyers (during and after aggregation) with a single or multiple sellers.
  • multiple sellers using c-INAs
  • the present system applies combinatorial auction processes to a unique commercial implementation of a multi-agent system.
  • Dynamic INAs (d-INAs) are double agents that switch roles from buyer to seller and vice-versa. D-INAs are used for arbitrage functions in which products, services and bundles are bought and sold at different locations for an immediate profit. In these instances, information currency is critical, so AAs are particularly important.
  • INAs interact with intelligent transaction agents in order to obtain information necessary to complete transactions.
  • ITAs interact with AAs in order to analyze and synthesize both general economic data and specific buyer/seller information. Once ITAs clear a transaction, for example, with a credit check or financing approval, the negotiation can be completed.
  • the interaction of specific time-sensitive functions occur in sequential order with the use of different appliances until the teams goal is completed; multiple functions may be processed simultaneously, with different orders at different times, so varied orders of completion will occur. In general, while maximum temporary efficiencies do exist, there is not necessarily a single way to prepare all of the projects to satisfy orders via the processing of specific operation sequences.
  • system architecture since the distributed system adapts to changing prices, the system architecture has self-organizing aspects similar to trading bazaars. Finally, the system architecture is designed to integrate into supply chain management (SCM), enterprise resource planning (ERP) and customer relationship management (CRM) software systems.
  • SCM supply chain management
  • ERP enterprise resource planning
  • CRM customer relationship management
  • CSAs also integrate with promotions and risk management options (RMOs) to invite customers with incentives such as time-sensitive promotional opportunities.
  • RMOs promotions and risk management options
  • This marketing integration mirrors how commercial systems actually work, but is missing in prior systems. For example, products may be bundled with services (financing, warranties, insurance, etc.), product features may be upgraded, or delivery time accelerated, in order to benefit unique buyer preferences. Further, proximity marketing is integrated with mobility in a MAS by providing time sensitive promotional opportunities to agents at a particular place. This advantage provides a bias to promoters that can use greater computation resources at their preferred location in order to maintain competitive advantages in negotiations.
  • AAs Analytical agents go beyond the typical pattern recognition and data mining tools.
  • EC evolutionary computation
  • AAs are powerful Al applications that inform and integrate with CSAs, lAs, INAs and ITAs.
  • AAs process complex data analyses and syntheses to increase system efficiencies.
  • AAs are the eyes of the system, while GP is the brain. Because they use Al and evolutionary computing processes, the system actually "thinks.” Consequently, AAs can anticipate market changes based on scenario forecasts.
  • the ability of AAs to adapt their programming to accommodate changing market situations is a critical step forward in research capabilities. This goes far beyond limit-order type securities program trading that previous computer exchange technologies have employed.
  • the negotiation-enabled MAS is intended to produce a computational system that mirrors the complex commercial psychology of markets. In essence, the system develops a process that emulates intuitive methods for commercial procurement. Thus, businesses and consumers can conduct commercial activities the way they prefer, namely, by employing direct contact approaches. For example, specific users employ regular patterns of commercial behavior. Hence, project-driven transactions can evolve into long-term business relationships.
  • the automation and efficiency aspects of the commercial negotiation and transaction aspects of the MAS increase value in the supply chain.
  • INAs provide disintermediated and automated negotiation in a distributed environment that emulates ordinary commercial relationships.
  • INAs use multivariate negotiation beyond merely price alone leading to a far more robust negotiating environment.
  • INAs promote competition between sellers, thus enhancing market efficiencies for buyers, by using simultaneous multilateral negotiation techniques.
  • INAs go beyond earlier systems not only because of their integration with AAs, CSAs and ITAs but because they use Al applications.
  • EC technologies - such as GP, GA and NN as well as integrated negotiation, auction and pricing strategy modules - INAs behave more independently than earlier systems.
  • Negotiation session parameter selection is enhanced by intelligent agents endowed with "judgment" for promoting optimal commercial trading processes. Such autonomy is particularly suited to the dynamism of the distributed database system.
  • C-INAs further empower customers by allowing both aggregation and complex multi-item sales.
  • Computational systems that use aggregation resemble multiplayer commercial markets because such processes use global information to benefit both buyers and sellers.
  • d-INAs allow intermediary-free arbitrage that facilitates complex shifting trading role-playing. Such applications represent a dramatic leap beyond current intermediated business-to-business exchanges.
  • Marketing and financial services are integrated into the system in novel ways. Marketing opportunities are integrated with showcases, AAs, CSAs, INAs and ITAs by accessing constantly updated promotion modules.
  • Marketing services include promotions, proximity and wireless marketing opportunities, RMOs, transaction contingencies and time- sensitive offers.
  • Financial services include accountability and credit reporting, banking offers and insurance, warranty and other risk analysis and risk limiting opportunities.
  • Al requires robust computation resources, which are provided as layered services. By cross-selling these dozens of specific services continuously in the distributed network environment, the system is flexible, scalable, pragmatic, integrative, self- organizing and effective. These services are sold in layers
  • agents operate in a distributed MAS with autonomy and mobility because they apply Al methods in a demand-initiated negotiation process.
  • Mobility has several advantages in the present system, including (1) communications failsafe in the event of interruption, (2) less cost because of enhanced efficiency, (3) reducing lags in negotiation by eliminating communication latency and (4) providing the neutrality of a level playing field between buyers and sellers in order to overcome bias.
  • the system provides both buyers and sellers with a shorter learning curve in making and processing transactions, as well as greater diversity of choice.
  • the system thereby promotes increasingly fair and efficient transactions.
  • the system provides increased market reach, increased efficiency and, consequently, tighter production cycles that contribute to reduced inventory.
  • the system streamlines the sellers own acquisitions and thereby reduces supply lags.
  • transactions are increasingly project-based, and supplier relations are increasingly flexible.
  • the whole supply chain functions more efficiently.
  • the present system causes little disruption to existing commercial systems because it emulates them in the computational sphere.
  • the system integrates well with current company ERP, CRM and SCM software systems. By automating such information exchange, negotiation, marketing and transaction processes, productivity rates are increased. Taken together, these advantages imply a sustainable competitive advantage for commercial sellers.
  • the system creates, promotes and enhances competition among sellers, making markets increasingly efficient for buyers. Buyer choices are increased and transactions costs diminished. Multilateral competition for a buyer in a distributed computational environment increases buyer efficiency and productivity while also diminishing transaction costs. Such a global computational sales and trading system allows increased vendor competition. This, in turn, promotes multi-item competition with minimized search costs.
  • Mobility eliminates negotiation bias that may otherwise limit operations to specific locations. Mobility also allows increased failsafe computation processes because the participants are relying less on (costly) communications systems that are prone to periodic failure.
  • the integration of marketing and financial services provides additional value to both buyer and seller. These services are fully integrated into the system.
  • the combination of promotions and risk management options offers a push-pull approach to market incentives in a distributed environment.
  • the use of AAs optimally leads to improved accuracy of information, particularly benefiting the activities of INAs and ITAs. The use of this information - both its analysis and timing - is critical to the development of sustainable competitive advantages.
  • 01/33464 A1 customer demand-initiated system and method for on-line information retrieval, interactive negotiation, procurement and exchange.
  • a reverse auction process is described in Godin, U.S. patent * 5890138.
  • a disintermediated auction system is described in Fisher et al., U.S. patent # 5905974. None of these approaches employs Al in a
  • Figures PA1, PA2, PA3, PA5 and PA6 outline these early models.
  • the first three systems involve centralized exchange processes. Buyers and sellers meet at a specified time and location to bid for items.
  • centralized exchanges are not always used, such as in the case of the Fishmarket model, brokers are used to intermediate the exchange of transactions.
  • Only the Contract Net protocol - which employs broker agents - is structured for distributed exchange processes in which the parties to the transactions are in diverse locations. All four main systems award contracts to victorious buyers on the basis of price. All four processes also use an ascending price buyer-side auction market model for competitive bidding and a seller-based approach to winner determination.
  • Automated agents are used in all systems primarily to mimic the behavior of buyers and sellers.
  • SWARM economic relationships are modeled. These main systems represent the early research in automated agent-based commercial systems.
  • EDI Electronic communications network
  • ECN electronic communications network
  • FIGS PA9A and PA9B describe an intermediated demand- initiated procurement system. This system is the first to emphasize buyer- driven (reverse-auction) commercial transactions, but is limited to intermediated exchange processes.
  • FIG. 1 illustrates a traditional approach to the integration of SCM, ERP and CRM software technologies in a supply chain. It uses the example of a big company that produces goods or services intermediating between small company sellers and buyers.
  • database management systems dbms
  • dbms database management systems
  • IBM database management systems
  • newer dbms involve the ordering of objects such as tables of data-sets.
  • Most contemporary dbms involve a fusion of object- relational (o-r) architectures.
  • o-r object- relational
  • Distributed search technology shown in figure PA12, was developed to respond to the shortcomings in the traditional search approach. Rather than collect data into a central depository, a distributed dbms searches numerous databases in real time. Assuming that the translation between the systems reveals compatibility of intercommunication, an initial query is sent to various databases, and search results are prioritized according to specified criteria for ordered display. Most search technology involving the Internet uses some combination of central and distributed technology approaches.
  • Fig. PA13 refers to a traditional aggregation system in which customers pool together in order to acquire a specific product.
  • the aggregation process allows a vendor to provide wholesale discount pricing, upon which, after a specified time, the buyers and seller ultimately agree.
  • the aggregation process automates a wholesaler intermediary function, in order to clear markets, but it has been increasingly challenged in the Internet age because supply chain layers are more easily eliminated and more information is available in a distributed computer network.
  • Fig. PA14 refers to a simple method of providing intermediated option contracts. This process allows a vendor to hedge a risk by selling an item to a second party even after it has initially agreed to sell it at a specified price to a first party.
  • the seller is obliged to pay the first party a pre-agreed penalty if the seller exercises a contingency to sell the item to another party, the seller can make more money - thus maximizing its benefits - by paying the penalty and selling to the second party for a greater profit than the first price plus the penalty.
  • finance and trading firms must utilize these risk management strategies in complex ever-changing markets so as to maximize revenue and optimize profits.
  • Fig. PA 1 is a schematic diagram of the operation of a multi-agent system.
  • Fig. PA 2 is a schematic diagram of an agent-based contracting system.
  • Fig. PA 3 is a schematic diagram of an agent marketplace.
  • Fig. PA 4A illustrates a simple system for integrated negotiation.
  • Fig. PA 4B shows a method for bilateral argumentation.
  • Fig. PA 5 is a schematic diagram of several stages of a contracting system.
  • Fig. PA 6 is a flow chart of the Contract Net Protocol.
  • Fig. PA 7 A is a schematic diagram of EDI as a paper replacement technique.
  • Fig PA 7 B is a schematic diagram of EDI as a process elimination technique.
  • Fig. PA 8 shows a simple electronic communications network (ECN) system.
  • ECN electronic communications network
  • Fig. PA 9 A and B are schematic diagrams of an intermediated demand- initiated procurement system.
  • Fig. PA 10 is a schematic diagram of a traditional supply chain with customer and supplier relationships.
  • Fig. PA 11 illustrates a traditional search method in an intermediated network system.
  • Fig. PA 12 illustrates a distributed search process in a disintermediated network system.
  • Fig. PA 13 illustrates a traditional aggregation method.
  • Fig. PA 14 illustrates a method for intermediated option contracts.
  • Fig. 1 is a schematic diagram showing the architecture of a cooperative communications network (CCN).
  • Fig 2 is a schematic diagram describing the relationships between the layers of a CCN system.
  • Fig. 3 is a schematic diagram of a showcase database system.
  • Fig. 4 illustrates multiple vertical databases.
  • Fig. 5 is a schematic diagram of a showcase database.
  • Fig. 6 is a schematic diagram of showcase database operation.
  • Fig. 7 is a schematic diagram of showcase data flow.
  • Fig. 8 is a flow diagram of the inter-agent system architecture.
  • Fig. 9 shows how rivers of data flows in a CCN operate.
  • Fig. 10 is a schematic diagram of commercial search agent (CSA) system architecture.
  • CSA commercial search agent
  • FIG. 11 is a flow diagram of a CSA first query.
  • Fig. 12 is a schematic diagram of CSAs indicating search priorities.
  • Fig. 13 is a schematic diagram of a CSA method used as an initial commercial search request.
  • Fig. 14 is a schematic diagram of CSA filtering methods.
  • Fig. 15 illustrates an apparatus for disseminating promotions in the system.
  • Fig. 16 is a schematic diagram illustrating proximity marketing for mobile INAs.
  • Fig. 17 shows the promotional discounting process.
  • Fig. 18 shows a dynamic pricing model with adaptive peak and off-peak pricing along a product or service cycle.
  • Fig. 19 shows a method for pricing discount promotions.
  • Fig. 20 illustrates promotions integrated with the CSA and Showcase database.
  • Fig. 21 is a schematic diagram illustrating a process of risk management option (RMO) contracts in a distributed network system.
  • Fig. 22 is a schematic diagram showing processes for transaction contingency logistics in a distributed contracting system.
  • Fig. 23 is a schematic diagram of information collaboration in a distributed network system architecture for use with made-to-order customization.
  • Fig. 24 is a schematic diagram illustrating a collaboration process for made-to-order (MTO) customization.
  • Fig. 25 is a schematic diagram of intelligent negotiation agent (INA) interactions in a multi-agent system with an emphasis on buyer agent (b-
  • INA INA
  • seller agents' s-INAs
  • Fig. 26 is a flow diagram representation of sequences of INA interactions.
  • Fig. 27 is a schematic diagram of a method for pre-negotiation in a multi- agent system.
  • Fig. 28 is a flow diagram representation of a method for time-based concealment of negotiation strategies in a distributed contracting system.
  • Fig. 29 is a schematic diagram of a system of INA logistics.
  • Fig. 30 is a flow diagram representation of a method for INA interaction.
  • Fig. 31 is a schematic diagram of a method for INA interaction.
  • Fig. 32 is a schematic diagram of an INA architecture emphasizing the initial interactions.
  • Fig. 33 is a flow diagram representation of an INA system architecture emphasizing negotiation interactions.
  • Fig. 34 is a schematic diagram of time-based negotiation sequences.
  • Fig. 35 is a schematic diagram of a method for initial INA mobile location protocol settlement.
  • Fig. 36 is a flow diagram representation of a tournament configuration of
  • Fig. 37 shows multivariate negotiation methods.
  • Fig. 38 shows automated negotiation sequences for item attributes with pre-established parameters.
  • Fig. 39 A and B is a schematic diagram of a demand-initiated automated negotiation in a distributed system illustrating mobility.
  • Fig. 40 shows multilateral distributed competition as a competitive double shout negotiation process.
  • Fig. 41 illustrates an INA negotiation module, including a schema of negotiation methods.
  • Fig. 42 illustrates an INA auction module, including a listing of several auction types.
  • Fig. 43 is a schematic diagram revealing the interactions of the INA negotiation module with the INA auction module.
  • Fig. 44 illustrates the pricing strategy module in the context of interactions with AAs.
  • Fig. 45 illustrates the interaction dynamics of INA "personalities”.
  • Fig. 46 is a schematic diagram of a system for the interaction of neutral cooperative INAs, including intermediation and aggregation applications.
  • Fig. 47 is a schematic diagram showing the sources of a C-INA transaction initiation.
  • Fig. 48 is a flow diagram representation of a method for B-C-INA based aggregation.
  • Fig. 49 illustrates several main automated aggregation category structures in a distributed network system.
  • Fig. 50 shows an INA based mass pooling approach to aggregation.
  • Fig. 51 illustrates a disintermediated aggregation method using C-INAs.
  • Fig. 52 is a schematic diagram of a disintermediated aggregation method for mass customization.
  • Fig. 53 is a flow diagram representation of a method for using dynamic INAs as double agents for arbitrage applications.
  • Fig. 54 illustrates an intermediated method for performing a combinatorial auction (CA) between a single seller and multiple buyers.
  • CA combinatorial auction
  • Fig. 55 illustrates a method for performing a CA with INAs between multiple buyers and multiple sellers in a single session.
  • Fig. 56 illustrates a method for winner determination of an interactive multi-lateral auction in a final session.
  • Fig. 57 is a flow diagram representation of a method for filtering variables for multi-item CAs.
  • Fig. 58 is a flow diagram of a disintermediated method of multi-item bidding from one seller to multiple buyers.
  • Fig. 59 is a flow diagram of a disintermediated method of multi-item bidding between multiple sellers and a single buyer.
  • Fig. 60 is a schematic diagram of a disintermediated method of multi-item aggregation of pre-set bundles between multiple buyers and multiple sellers.
  • Fig. 61 is a schematic diagram of a disintermediated method of multi-item bidding between multiple buyers and multiple sellers.
  • Fig. 62 A and B is a flow diagram representation of a disintermediated method of aggregation whereby multi-item bundles are exchanged between multiple buyers and multiple sellers using c-INAs.
  • Fig. 63 is a flow diagram of a disintermediated method for conducting arbitrage of multi-item bundles between multiple sellers and multiple buyers using d-INAs.
  • Fig. 64 illustrates multi-factorial bidding approaches by listing item variables that can be sorted.
  • Fig. 65 shows a multi-feature example of item-factors that can be sorted in the case of personal computer configurations.
  • Fig. 66 shows examples of multi-item bundle category applications.
  • Fig. 67 is a schematic diagram representation of a seller mobile transaction agent (s-ITA) system architecture.
  • Fig. 68 is a flow diagram of s-ITA operation.
  • Fig. 69 is a schematic diagram of a s-ITA and b-ITA system process in the final negotiation with one seller.
  • Fig. 70 shows the ITA service categories, including the buyer and seller roles.
  • Fig. 71 describes ITA service categories.
  • Fig. 72 is a flow diagram representation of b-INA micro-agents generated particularly for a negotiation session in a mobile application.
  • Fig. 73 is a flow diagram representation of a method for genetic algorithms to be applied to multi-agents system.
  • Fig. 74 is a flow diagram representation of a method for neural networks applied to a multi-agent system.
  • Fig. 75 is a flow diagram of a genetic programming system process.
  • Fig. 76 is a schematic diagram representation of methods for a genetic programming learning schemas.
  • Fig. 77 is a flow diagram representation of a method showing evolutionary computation applications to autonomous agents.
  • Fig. 78 is a schematic diagram showing Al applied to agency in a distributed system.
  • Fig. 79 is a flow diagram of an evolutionary computation architecture with AA and INA applications.
  • Fig. 80 is a flow diagram illustrating layers Al for optimum agent mobility.
  • Fig. 81 is a schematic diagram showing AA architecture from buyer and seller viewpoints.
  • Fig. 82 is a schematic diagram showing kinds of data analysis and synthesis.
  • Fig. 83 is a schematic diagram representation of an AA data flow process.
  • Fig. 84 is a schematic diagram representation of methods of data mining, with emphasis on interactions between a commercial search agent (CSA) and AA.
  • Fig. 85 is a flow diagram representation of methods for collaborative ; filtering for cross-marketing recommendations applications.
  • Fig. 86 is a schematic diagram of b-AA operation with mobility.
  • Fig. 87 lists the variables of a super-score system in a multi-agent system.
  • Fig. 88 lists the variables of market and economic analytics in a multi- agent system.
  • Fig. 89 lists the variables of an accountability index in a multi-agent system.
  • Fig. 90 lists the variables of a financial criteria index in a multi-agent system.
  • Fig. 91 lists insurance risk factors for use in a multi-agent system.
  • Fig. 92 lists service categories in a multi-agent system.
  • the system represented by the present invention has numerous distinctive embodiments.
  • the present disclosures illustrate in detail the main ideas of the invention and are not intended to restrict the invention to a single embodiment.
  • the system and methods incorporated in the present invention are implemented by using software as applied to networks of computers, microprocessors or mobile computers.
  • Software is stored in memory on computer disk drives.
  • the microprocessors in the computer hardware use database software to store data used by software applications such as intelligent software agents.
  • Agents are computer software program code that can be activated to perform specific functions. Once activated, agents can be executed in a node of a computer network, or can move from node to node to manifest mobility.
  • the present invention aims to solve problems in the area of computation for automating demand-initiated sales processes in a distributed network.
  • the present invention uses a distributed database system to automatically store data about goods and services, and access, analyze, and collaborate about the data.
  • the present invention further discloses a demand-initiated sales application of intelligent software agents in a multi-agent system (MAS).
  • MAS multi-agent system
  • the agents perform an automated, multivariate negotiation for individual items in a multilateral interactive environment between a buyer and at least two sellers.
  • the system further uses intelligent negotiation agents (INAs) to perform aggregation, arbitrage and combinatorial functions in a MAS.
  • INAs intelligent negotiation agents
  • ITAs intelligent transaction agents
  • One key feature of the software agents in a distributed MAS is the use of mobility. Such a computer system applies the frontiers of artificial intelligence.
  • a showcase is a database that contains constantly updated information from a corporate database.
  • Corporate databases (1005) receive data inputs in real time as seller inter-agents (S-IAs) (1015), filter information for seller showcases (1045).
  • Analytical agents (AAs) for both seller (1010) and buyer (1013) analyze and filter information: seller AAs (S-AAs) analyze market data inputs for seller inter-agents (S-IAs) while B-AAs analyze market data inputs for B-IAs.
  • Data used can be objects, codes, text, images, multimedia or other formats.
  • RMO promotions and risk management option
  • S-IAs promotions and risk management option (RMO) contracts are provided to the showcase by the S-IAs from promotion and RMO modules (1020).
  • RMO risk management option
  • buyers use B-IAs by setting the parameters of software preferences at the customer graphic user interface (GUI) (1040).
  • GUI customer graphic user interface
  • B-IAs are intermediary software programs that interface with the customer GUI and other agent programs to constantly re-adjust the parameters of automated software agents.
  • buyer and sellers can bypass the showcase system and collaborate (1035) B-IA and S-IA interactions in order to mutually identify item specifications prior to negotiation sessions.
  • the CSA can access showcase databases for a first search query and then proceed to a negotiation session.
  • a B-AA has analyzed a product or service category
  • this information is forwarded to a B-IA, which then initiates a search of showcases by activating a commercial search agent (CSA) (1060).
  • CSA commercial search agent
  • the B-IA can use the search results to interact with S- lAs in a pre-negotiation session (1055), which sets the rules of encounter for negotiation sessions between the buyer and two or more sellers.
  • the Buyer Intelligent Negotiation Agent (B-INA) (1085 and 1090) enters into interactive multi-lateral negotiations with at least two seller intelligent negotiation agents (S-INAs) (1075 and 1080).
  • S-INAs use pricing, negotiation, and auction modules to automate the negotiation interactions with the B-INA.
  • B-ITA buyer intelligent transaction agent
  • B-ITA buyer intelligent transaction agent
  • the B-ITA must clear the transaction for the B-INA by double-checking and clearing all terms of the S-INA. Once all terms are mutually agreed upon, the transaction is concluded and all inter-INA activity terminates. ITAs in turn update their respective AAs with data from the transaction.
  • Fig. 2 describes the CCN system layers and their relationships.
  • the layers of a CCN can be described as follows: (1) Corporate databases and customer GUI are the front end or top layer; (2) showcase databases are at the second layer; (3) AAs, both buyer and seller, as well as promotions and RMOs, are at the third layer, which represents analytics; (4) Buyer and seller inter-agents and collaboration are in the middle, interagency layer, at layer four; (5) CSAs are at the fifth (search) layer; (6) Several varieties of INAs are at the sixth (negotiation) layer, and; (7) Buyer and seller ITAs are at the seventh (transaction) layer.
  • This distributed database infrastructure and this integrative multi-agent system differs substantially from prior systems, as does the system architecture described in these figures.
  • the specific methods for actions and interactions also vary from earlier approaches.
  • the MAS is integrated into the database system; the database system configuration provides the forum for the MAS.
  • Fig. 3 illustrates a showcase database system whereby corporate databases (1155) are using analyses from S-AAs (1180) as well as receiving corporate data inputs.
  • S-IAs (1157) use S-AAs to analyze and filter data in order to continuously place new products and services in each respective company showcase database (1160). Similarly, showcase data is continuously purged to reflect the most current commercial activities.
  • S-IAs also constantly receive data from promotion and RMO modules (1165) (also informed by S-AAs). Showcases thus reflect the most current data on products and services, and are informed by market data; as market data changes, the showcase data constantly updates to respond.
  • Showcases typically operate in the context of specific industries. See figure 4. Showcases are more focused in vertical industries and tend to share common specialized languages (both for items and the codes
  • Each vertical showcase system is similar to the yellow pages, but for worldwide access to an industry.
  • a CCN can overlap in more than one industry.
  • Such horizontal (or trans-vertical) showcases combine multiple categories of items and can be accessed by multiple showcase systems.
  • a CCN can be customized to a specific business, such as a very large corporation that may use thousands of vendors or customers.
  • Such a showcase database system has advantages over other systems.
  • this CCN architecture is scalable. Seventh, the system is structured to be omni-directional and thus accessible from anywhere.
  • promotions (and RMOs) are integrated into the showcase system, marketing (push and pull) is key to its pragmatic operation.
  • Fig. 5 illustrates how S-AAs feed filtered data to S-IAs (1325) for inclusion into a showcase.
  • S-AAs (1310) filter data from market sources (1300) to a showcase (1345).
  • data included in showcases is primarily derived from corporate databases (1320), data is also input from the pricing module (1330), promotion module (1335) and RMO module (1340). Further description of these modules are made in figures 14-22 and 41-44.
  • the showcase database view will be different since it represents snapshots of constantly moving data streams that can be seen over time (1315).
  • a showcase can be accessed from various locations (1350) by CSAs.
  • S-IAs As illustrated in Fig. 6, corporate databases (1340) provide raw data to S-IAs (1344).
  • S-IAs receive data from S-AAs (1346) that filter raw market data from various sources (1342) as well as promotion and RMOs (1350).
  • the S-IA continuously filters the data in real time to the showcase database (1348).
  • the showcase has two main components: (1) a database for specific items and (pre-set) bundles and (2) a database for customizable items in the collaboration process. Since data in the showcase is time sensitive, such data is constantly renewed and old data is purged.
  • the CSA (1352) can access the showcase for the initial query in order to initiate the negotiation process.
  • Fig. 7 shows the flow of data from the corporate database and S-
  • Inter-agents are used as mobile intermediaries between various agents and databases (or GUIs). lAs are categorized as either buyer-side or seller-side. lAs also interact with each other. Fig. 8 describes inter-agent system architecture. On the seller side, a seller (1370) interfaces with a corporate database (1374) via a GUI
  • the seller inter-agent (S-IA) (1382) intermediates between the corporate database and the showcase(s) (1390) as well as the S-AA
  • the buyer inter-agent (B-IA)(1384) intermediates between the customer-side GUI (1378) and the CSA (1388) and between the GUI and the B-AA (1386).
  • lAs can be used more expansively to include interaction with INAs, ITAs and market data. Because they are intermediaries, lAs may be mobile. Their locations change in sequence or they may alternate.
  • Fig. 9 illustrates the rivers of data flows in a CCN system. This figure shows the seller and buyer sides as well as the top (database and inter-agent) layers and the bottom (negotiation and transaction processing) layers. Note that most actions in this view of the CCN system involve interactive (i.e., not unidirectional) functions. This view also downplays the primary negotiation functions to emphasize the supporting structure.
  • FIG. 10 shows the sequence of CSA actions in the first query.
  • Market data (1552) is input into B-AAs (1620).
  • Various promotions (1585, 1590, etc.) are input into showcases via S-IAs.
  • a CSA (1625) accesses a B-AA and, with data from the AA, the CSA accesses the showcases, which have been informed by promotions, in an orderly sequence. The showcases are constantly updated by the promotional modules.
  • the CSA filters the best promotions and then accesses selected showcases, which respond by providing specific data.
  • Fig. 84 shows CSA and AA interactions with an emphasis on data mining approaches.
  • the CSA asks a specific question in order to receive specific commercial data from a showcase. Key words of the first CSA query are ordered and ranked so as to provide a feedback in a particular sequence.
  • Fig. 12 shows the flow of sequences involving the CSAs search method.
  • the CSA then accesses (1740) various showcases in a CCN (or in more CCNs) according to registry code sequences.
  • the showcases are accessed according to a time-sequence synchronization that provides code-priority to specific showcases based on factors upon which CCN coop members can agree.
  • An efficiency optimization process calibrates the synchronization access system.
  • priorities that can be conferred on a CSA include popularity, specialty, price, quality, etc.
  • the CSA sends out a code seeking a match to a request for specific attributes to the showcases in a CCN (1750).
  • the showcase database(s) Upon receipt of the code request, the showcase database(s) respond (1760) by sending back matching relevant query attributes.
  • the CSA orders and ranks responses according to configured preferences (1762).
  • the CSA registers the search results to the buyer GUI (1764) or to B-IAs or B-INAs.
  • the CSA (1805) while being informed by a B-AA (1800), requests information for initial item data from each of several showcases.
  • showcase 3 allows the customers to buy the item at an initial price.
  • the buyer can negotiate with at least two (1820 and 1825) sellers by proceeding to a pre-negotiation session (1830), which sets the terms of the negotiation process.
  • Direct interactive bilateral negotiation proceeds between a B-
  • the negotiation session proceeds, but in the case of S-INA1 , the buyer may choose to buy the item at an initial price (1860) and proceed to closing the transaction with an ITA (1880).
  • the CSA represents the first request for information, including item features, terms and price.
  • the CSA can filter information about commercial items according to various categories including price, location, item niche, availability, bundles, accountability, and past experiences.
  • Fig. 14 illustrates these CSA filters (1910).
  • the accuracy and quality of data are also filtered and purified (1965). These filters are manually configurable so as to track the specific categories, either individually or in combination.
  • sellers can invite specific buyers to special promotions by micro-casting unique discount or incentive opportunities (1920). Taken as a whole, these seller-driven inducements create a "demand-shaping" of distinct buyer needs (1930). By influencing buyer activity, the seller can control its own suppliers more smoothly and even out buyer demand. Such demand shaping occurs at the CSA and promotional level.
  • the promotional (& RMO) data is used by buyers to generate choices (1970) for negotiation sessions.
  • Fig. 15 shows that a customer can detail preferences (2015) at the CSA after obtaining general item data from a broadcast promotion (2010). The customer then accesses the
  • CSA (2025), which, in turn, accesses the showcases (2030) in a specific industry. Showcases can receive targeted ads (2035) based on pre- specified preferences, which are accessed by a next data stream of the CSA. Finally, the CSA makes a buyer request for sellers' information (2020).
  • a B-INA moves to another location (2042).
  • An S-INA accesses a database via an S-AA to determine whether the B-INA qualifies for a specific promotion (2046) at that time. If the B-INA does not qualify, no promotion is provided (2048). On the other hand, if the B-INA does qualify, the S-INA offers it a specific, time-sensitive promotion at a specific location (2050) by accessing the promotion module (2044).
  • These promotions can be in the form of "exploding" offers that diminish over time until a deadline is realized and they expire.
  • Proximity marketing allows sellers to shape demand by influencing buyer behavior with incentives. These incentives may be induced by the seller because it receives unexpected reductions from its suppliers. The changed circumstances induces the seller to offer improved terms by shifting opportunity to other, more available, supplies to buyers. Proximity marketing also allows sellers to exploit opportunities by immediately offering promotions for qualified buyers in the context of a demand- initiated sales system. Such a process may confer an advantage on a particular seller.
  • One of the advantages of proximity marketing is that sellers can offer an incentive to a B-INA at a specific place in order to induce the B-INA to conduct negotiations at the. home of an S-INA. The location priority would confer an advantage on the S-INA by providing easy access to computation resources.
  • Such proximity marketing with mobility in a distributed MAS not only enhances revenue management but also profit maximization through increased efficiencies.
  • Fig. 17 There are different main types of promotions as outlined in Fig. 17. In general, products that use decaying technology (2060) and services that use off-peak capacity (2070) are discounted. See also Fig. 18. In addition, items that are bundled, possess fewer features or lower quality, have multiple units or are less time-sensitive (2080), are discounted. Information about these discounts is broadcasted in the form of promotions outlining these sales opportunities.
  • a Dynamic pricing curve is illustrated in Fig. 18. This figure shows the decline of prices related to off-peak services or to trailing edge technology.
  • Fig. 19 shows a list of promotional categories to which item prices tend to fall. These categories are thus used in marketing promotions by the promotion module.
  • Promotions are typically included at the front end ofthe system, at the CSA or showcase level. As Fig. 20 illustrates, promotions can come in the form of ads or inducements. Ads are pushed as broadcasts or targeted as customized. Inducements, on the other hand, are invited and driven by the seller, or intentionally requested by a buyer. Discounts are again specified according to the product decay curve (or by supply and demand curves) or services yield management curve. It is important to note the multivariate nature of discounting beyond the price factor alone, because quantity, quality, features, bundles, and time-sensitivity are each important criteria that affect customer demand. In fact, these multi- faceted item criteria allow promotional cross- marketing for sellers to provide marketing opportunities before or during an initial search.
  • RMO Risk management option contracts represent a seller- induced opportunity to control risks by offering buyers contract clauses that penalize sellers. While promotions are pushed to buyers from sellers, RMOs are "pulled" by buyers from sellers. These contingencies allow sellers to hedge their bets on items that may not be fully in their control. On the other side, buyers can exploit these risks as opportunities, either to get good value, or to receive a penalty from the seller in the event that they must withdraw from a contract by exercising a contingency clause. Because the terms of RMOs must be agreed to early in the transaction, they are on the level of a promotion. Further, like promotions, since the underlying circumstances for sellers constantly change, RMOs conditions and contingencies constantly change. Therefore, as a CSA accesses a showcase, the most recent promotional and RMO data is constantly updated. Since showcases are structured in a distributed system, RMOs are similarly distributed.
  • Fig. 21 outlines the operation of RMO contracts in a distributed system.
  • S1 agrees to sell item(s) at specified terms to a customer (C1), while S1 agrees to pay a specified penalty to C1 if it cannot comply with the terms of the transaction.
  • Seller AAs (2235) are constantly receiving market data inputs (2215) and dynamic pricing data inputs while assessing risk priorities and preferences (2245). Being an opportunist and rational agent, S1 constantly seeks a better deal than he arrived at with C1.
  • S1 finds C2 and enters into a new arrangement with different terms. S1 then commits to buying the item from S2 in order to re-sell to C2, according to optimal terms.
  • S1 sells the item(s) to C2 (2240) at a higher price than original terms arranged with C1; thereby, at (2265), C1 receives a penalty from S1; either the item is merely delayed, which may warrant a mild penalty (2295) or cancelled, which will typically justify a more significant penalty (2300).
  • Risks are either re-packaged (2260) or time shifted (2290). If they are re-packaged, the risks can be either concentrated (2255) (for maximum return on the upside) or diluted (2310) for risk diversification (minimum loss on the downside). By time shifting various risk contingencies, risks can be diluted over time so as to overcome a temporary supply imbalance. Risk penalties can come in the form of cash or future discounts (coupons).
  • RMOs allow sellers flexibility, especially from unforeseen supplier circumstances, but also provide buyers with distinct market opportunities. RMOs function as a sort of risk arbitrage by shifting risk in unforeseen circumstances.
  • the mobility aspect ofthe present system allows unique advantages in a distributed network because sellers are able to more accurately and immediately respond to complex market circumstances up and down the supply chain.
  • RMOs when combined with promotions, provide a powerful marketing integration component to a transaction system.
  • a reverse RMO may be used as a form of performance bonus to reward a seller for an excellent job such as accelerated contract terms favorable to a buyer.
  • a compounded RMO may also be used by D-INAs for arbitrage applications. Such complex RMOs are used by the original seller to the D-INA in its initial role as buyer and then by the D-INA (in its secondary role as seller) to a buyer. As an intermediary, the D-INA effectively spreads the risk from first buyer to final seller. Agents do not only have the power to negotiate and contract, but also to use contingencies, by both buyers and sellers and with penalties and without penalties (depending on the nature of the contingencies). In Fig. 22, the logistics of transaction contingencies are shown.
  • a buyer may delay a seller negotiation for a time (2319).
  • One way to do this is for a buyer and seller to express an interest in a transaction (2320).
  • a B- ITA seeks to close the transaction (2323).
  • the buyer can end the transaction by withdrawing (2335) and seek to negotiate for other items (2337) given buyer priorities.
  • the seller may withdraw without penalty
  • the buyer and seller then agree to enter into a transaction (2321 and 2324). If the seller withdraws from the transaction, the seller activates an RMO contingency (2326) and the seller pays a penalty (2327). If the buyer contingency is not performed (2325), the buyer may withdraw with a penalty (2329). In this case, the buyer may return to the negotiation session and seek his second best choice (2337).
  • RMO contingency 2326)
  • the buyer may withdraw with a penalty (2329). In this case, the buyer may return to the negotiation session and seek his second best choice (2337).
  • JIT Just-in-time
  • MTO made- to-order
  • the information collaboration for MTO items is integrated with the CSA during an initial search, as well as before and after a search.
  • Collaboration can occur before the CSA's first query (by a B-IA and S-IAs), and after the
  • Fig. 23 shows the initial collaboration between a B-IA and at least two S-IAs.
  • AAs inform both buyer (2365) and seller inter-agents (2370 and 2385).
  • the buyer via the inter-agent, accesses the CSA (2400) in order to get pricing and transaction term data from sellers.
  • the initial collaboration process bypasses a pre-set item showcase database input, and precedes a search query.
  • CSA search (2400) on the other hand, collaboration occurs by the CSA interacting with S-IAs (2420 and 2430) and a B-IA
  • a B-IA (2509), after being informed by a B-AA (2499), interacts with at least two S-IAs.
  • the S-IAs (2503, 2504 and 2505), after being informed by S-AAs (2495, 2496 and 2497) (which interact directly with corporate databases (2490, 2491 and 2492)), interact with the B-IA
  • collaboration can occur during a search. In another embodiment, collaboration can occur after a search but before a pre-negotiation session. INAs
  • Intelligent negotiation agents are complex autonomous software agents programmed to conduct interactive negotiations with specific rules or goals. Because INAs are intelligent - they use artificial intelligence technologies - they evolve their operations beyond initial pre- programmed parameters so as to adapt to changing market conditions. In order to develop adaptive programming, the agents operate within a multi- agent system (MAS) according to second-order system rules that govern the primary rules of the immediate negotiation functions. Such MAS metarules allow the agents to evolve operational negotiation rules in a complex distributed computer system. INAs operate in the context of buyer and seller interactions.
  • MAS multi- agent system
  • INAs have several main types, including seller-INA (S-INA), buyer-INA (B-INA), dynamic-INA (D-INA) that switches roles from buyer to seller and vice- versa, and cooperative INA (C-INA) [buyer C-INA (B-C-INA), seller C-INA (S-C-INA), lead B-C-INA and neutral C-INA] involved with aggregation and combinatorial auctions.
  • S-INA seller-INA
  • B-INA buyer-INA
  • D-INA dynamic-INA
  • C-INA cooperative INA
  • B-C-INA seller C-INA
  • S-C-INA seller C-INA
  • the present invention uses INAs in a distinctive demand-initiated system in which a B-INA, after performing an initial query with a CSA and receiving a report of initial bids from at least two sellers for a comparable item (or items in a bundle), bids simultaneously with the S-INAs.
  • the interactive negotiation process operates by the INAs using complex program code that (1) performs specific negotiation functions using game theoretic processes, negotiation strategies and dynamic pricing information and (2) evolves beyond the initial negotiation parameters in order to conduct multivariate multilateral anticipatory realtime interactive negotiations with mobility in a distributed computation system by employing Al technologies.
  • the demand-initiated buyer driven negotiation process operates by each agent (a) receiving, (b) reviewing and (c) evaluating the input bids and (d) deciding the best options to respond. While an expert system preprogrammed with negotiation parameters can perform these operations, much like a sophisticated chess game, the present system applies Al technologies to adapt the negotiation operations beyond the initial parameters (within second-order MAS meta-rules).
  • Such Al-applied adaptation beyond initial negotiation parameters allows a B-INA to assess incoming bids, evaluate the bids and chose an optimal response.
  • the evaluation process consists of comparing each respective bid to various scenarios within programmed parameters.
  • INAs access AAs
  • a B-INA interacts with at least two S-INAs in order to agree to multi-lateral rules of negotiation (at 2585).
  • the agents proceed to negotiation sessions (2590) between the B-INA and S-INA1 and between the B-INA and S-INA2.
  • the locations ofthe interactive negotiations may be at the buyers, the seller(s)', or alternate between the two types of locations. At some point, the buyer selects a winner (2595) and awards a contract.
  • the winning S-INA then passes the transaction terms to the winning sellers S-ITA (2605) and the B-INA passes the deal terms to its B-ITA (2600). If terms are not verified or certified (i.e. by credit checks), the transaction is then sent back to negotiation sessions for re-negotiation of those terms and then returned to the ITAs for clearance. Once the ITAs clear the transaction, the negotiation ceases (2610) and the session ends.
  • At least four sellers will be ranked at the first search response. See Fig. 36.
  • the (at least) four S-INAs will interact in the pre-negotiation session with the B-INA.
  • the negotiation session allows the B-INA to select two S-INAs on which to focus further negotiations (suspending, or stopping the others).
  • Fig. 25 emphasizes the negotiation process during the phases of negotiation once two finalist S-INAs have been selected for negotiation by the B-INA. Consequently, the winner is determined as referenced in the paragraph above.
  • a tournament configuration is present in which "contestants" are inevitably eliminated until a single winner is determined.
  • more than one seller can be selected for a transaction.
  • the B-INA may be in at least two different places at the same time.
  • the B-INA may alternate functions by rapidly moving from location to location, though this solution leaves the necessary problem of time-delay lags in activity.
  • Fig. 72 shows micro-agents.
  • the negotiation process between a B-INA and S-INAs may include counter- bidding processes directly between the B-INA and specific interactive S-INAs.
  • Fig. 26 illustrates a final stage of negotiation between at least two S-INAs and a B-INA. Without showing the pre-negotiation phase, or inter-agent activity, this figure describes the counter-offer process, at 2685 and 2695 between the B-INA and S-INA1 and at 2705 and 2710 between the B-INA and S-INA2.
  • the interactive counter bidding process may continue between buyer and seller INAs for multiple sessions. Since counter-bidding is based on factors beyond price alone, the potential criteria are compounded in complexity. Consequently, the process of negotiation with each seller may be protracted. In addition, since there are at least two sellers, a competition between the two bidders creates a sustained process of bidding not duplicated in single bidder type negotiation sessions. For instance, the competitive frontier of a negotiation is more likely to be extended and optimized with two or more seller bidders. The buyer has the option of disclosing all or part of the negotiation sessions with other S-INAs and such disclosures can act to increase competitive bidding.
  • INA may elect to either negotiate simultaneously with several S-INAs, or may prefer to narrow the field, as illustrated in Fig. 26, to two competitors, before selecting a final winner. After a winner is determined, sending messages to the remaining negotiators terminates negotiations with other S-INAs.
  • the pre-negotiation process is described. After a customer requests negotiation terms (2755) through a B-IA (2760), the customer proceeds to a B-INA (2765). Several sellers, shown here at 2770, 2775, and 2780, are selected and meet the B-INA at a pre- negotiation session (2785) to determine interactive negotiation parameters. These parameter factors include locations, protocols, auction methods, etc. If agreement cannot be reached with a seller, the B-INA may proceed to negotiate with another seller. If at least two sellers can agree with the B-INA about preliminary protocols, communication aspects and other meta-issues, the B-INA and S-INAs proceed to establish (at
  • One ofthe ways to conceal agent negotiation strategies is to use time based modulation to disguise agent interest.
  • Negotiation responses can be quick or slow depending on the intentions of the agents.
  • providing contradictory actions can conceal agent intentions.
  • disguised actions provide signals that are difficult for opposing parties to read.
  • One way to accomplish this is for INAs to employ a randomizer that can alter the composition ofthe content of a bid so as to deceive an opponents anticipation of moves.
  • INA logistics are described. After initiating the session (2690), agents are generated and identified by codes (2695).
  • the initial agent interaction protocols are generated (2970) in order for the agents to establish a common communication methodology.
  • Such communication processes involves translation (2975) and synchronization (2980). Failure to synchronize communication leads to a termination at 2995.
  • INAs may construct unique negotiation strategies using Al (2990) through an AAs information (2985). At this point, agents signal the intention (at 3000) to negotiate with other agents. After signaling to other agents, INAs send out communication streams (3005) to their home base, thereby constantly revealing to the home base their locations, status and plans as well as receiving periodic parameter modification updates from home.
  • a B-INA and S-INAs enter pre-negotiation sessions to set rules for further negotiations. Information about these pre-negotiation sessions is sent home (back to 3005). After pre-negotiation, a B-INA launches micro-agents (3015) in order to negotiate simultaneously with S-
  • INAs at different locations.
  • INAs enter the negotiation sessions, which can lead to agreement between a B-INA and a S-INA (at 3024) and a provisional transaction completion (3025). It can also lead to INAs' ceasing negotiation (at 3023) in which case INA settings are saved (3030) for later re-launch.
  • the S-ITA (3035) and B-ITA (3040) activate. Either ITA may return the transaction back to negotiations, or if both approve it, the transaction may close (3050) and agents self- terminate (3045) by saving INA settings for later re-launch (3030), and the session closes.
  • INA interaction sequences are described.
  • Fig. 30 after initiating an initial commercial search request (3075) by a CSA, agents pre-negotiate at 3080. Those that do not successfully complete pre-negotiation return to 3075.
  • agents Upon pre-negotiation completion, agents initiate negotiation sessions at 3082.
  • agents activate specific negotiation strategy and tactical modules (3085). While negotiating at different locations, buyers and sellers involved in the negotiations track mobile agents (3090) and continuously register the interaction activity with their home bases. Once a winner is determined (3095) by a buyer selecting a seller, negotiations between a buyer and seller lead to an initial commitment (3100).
  • a buyer or seller can withdraw from the initial agreement (3105) by exercising a contingency and return to an initial CSA request.
  • the successful INAs can push the transaction to the ITAs (3110); if the transaction is not completed by either a B-ITA or S-ITA, it is returned to negotiation sessions (3082). After the deal is finally closed, the settings are saved (3115), the agents self-disable, and the session closes.
  • INAs access AAs (3125 and 3130), which are both endowed by Al (3122), and then enter pre-negotiations (3135).
  • S-INAs access the pricing module (3145), negotiation strategy and tactical modules (3155), and the auction module (3160) before proceeding to the negotiation session(s) (3165).
  • B-INAs access the negotiation module (3155) and auction modules (3160) before proceeding to negotiation session(s) (3165).
  • the transaction continues to the ITAs (3170) and then either back to negotiations, or to close (3175).
  • the transaction can return from the negotiation session(s) to the pre-negotiation session(s) stimulated by either the buyer, or the seller agents.
  • Fig. 32 describes the early interactions
  • Fig. 33 emphasizes the sequencing of INA negotiation sessions.
  • Each showcase receives inputs from S-IAs, S-AAs and corporate databases.
  • Promotions and RMOs also interact with showcases (and B-INAs).
  • CSA CSA
  • CSA CSA
  • results displays results 3282
  • B-INA a B-INA
  • the B-INA proceeds at 3300 to the pre-negotiation sessions(s) with only two of the S-INAs (3295 and 3305). In the embodiment illustrated here, the field has narrowed from three to two.
  • four or more showcases can be accessed and at least four S-INAs can be involved in pre-negotiations and in negotiations, with a narrowing of the field from at least four to at least two until a winner S-INA is selected (at 3315) by a B-INA.
  • negotiation sessions are illustrated.
  • session one at least the three sellers are specified, but in session two (3443) negotiation occurs only between the B-INA and S-INA 1, on the one hand, and the B-INA and S-INA 3, on the other.
  • negotiation session three In negotiation session three
  • the B-INA focuses only on S-INA 3. After terms are negotiated and agreed to, the B-INA selects a winner and either proceeds to ITAs for completion or returns to 3440 for negotiation with several S-INAs. These negotiations may occur at various, or even alternating, locations. Each negotiating session can continue for numerous interaction sequences and may include criteria beyond price alone. The negotiation sessions may occur in sequences that narrow the field, as described in figures 32 and 33, or may occur simultaneously until a B-INA selects a winner.
  • Fig. 34 shows the time-based sequences of negotiation session(s).
  • a first search request leads to a first seller ask (3510) (or the first request information display) and to the first buyer offer (3515).
  • the first seller counter offer occurs at 3520 followed by a second buyer counter offer (3525) and second seller counter offer (3530).
  • the buyer may accept the second seller counter offer at 3535.
  • pre-negotiation session(s) (3570) determine the buyer (3575) and seller (3580) locations at which to conduct negotiation activities.
  • the B-INA can negotiate at its home or the sellers' home(s) or can alternate between locations at various times during the negotiation process.
  • Fig. 36 shows the narrowing process of INA winner determination in a tournament configuration.
  • four S-INAs are accessed at
  • Fig. 37 multivariate negotiation is described.
  • two S-INAs provide first "ask" information (3710 and 3715).
  • This first ask can be in the form of a price or, additionally, of a range of information about item features and qualities.
  • the existence of item information beyond price alone suggests that the initial search request is substantially more than a mere RFQ, which focuses only on price.
  • Such broader search request and response is also more conducive to custom orders.
  • the buyer provides a first counter-bid to each S-INA initial ask; each counter-bid can reference item features, quality, etc. as well as price.
  • the S-INAs provide their respective counter offers to the buyer first counter bid.
  • Each S-INA has access to customer and market data, supplied by S-AAs at 3720 and 3725, respectively.
  • each S-AA may supply or emphasize different kinds of data, second variable.
  • the seller provides a first price (3770), which is countered by a buyer (3772).
  • the negotiation proceeds to a final compromise price (3786).
  • the second set of variables is similarly negotiation by pre- established parameters until a final compromise is reached (3806).
  • the outcome of the second variable(s) negotiation may influence the first variables negotiation outcome, and thus the first variable may require negotiation.
  • the negotiation process is completed.
  • the negotiation between a B-INA and at least two S-INAs shows the complex dynamics of automated negotiation dynamics over multiple variables.
  • this multi-lateral multivariate automated negotiation process occurs with mobility in alternating locations with INAs moving program code as illustrated in Figs. 39A and 39B.
  • Figs. 39A and 39B illustrate the negotiation process in a distributed system with mobility between a buyer and seller.
  • the present example focuses on a one-to-one negotiation between a B-INA and a S-INA.
  • a buyer B-INA
  • S-INA seller
  • the INAs identify possible locations (3822) and specify agreed locations (3825) at which to negotiate.
  • the B- INA moves to the S-INA location with program code (3827).
  • the S-INA identifies incoming B-INA entry after activation and security protocol approval (3830) at the S-INA location.
  • the agents engage in (3832) and complete (3835) negotiation tasks, after which the B-INA "phones home” by notifying the buyer “home” computer of remote location activities by sending a message (3840).
  • the B-INA (3845) either terminates (or returns home) (3850) or assesses additional tasks using internal database and analysis (3855), assess (3857) and identifies (3860) the next location for task execution and moves to another locations (3865).
  • the B-INA After moving its program code (3870), the B-INA identifies a need for Al computation (3875), requests Al computation resources at a specified location (3880), identifies available Al computation resources
  • the B-INA receives (3895) and tests (3900) the Al computation resources at a specific negotiation sit ⁇ (3895). The negotiations are completed at the remote location (3905) and the B-INA returns home (3910).
  • the B-INA sends its "children" or micro-agents (cf. Fig. 72) to remote locations because it must be split into at least two parts in order to negotiate simultaneously with two S-INAs.
  • a B-INA (or its micro-agents) may negotiate simultaneously with at least two S-INAs at two or more S- INA locations (as well as buyer or intermediary locations) in an additional embodiment.
  • the B-INA and S-INAs may also alternate between the various locations according to the agreed negotiation requirements ofthe INAs.
  • Fig. 40 illustrates how a B-INA can simultaneously negotiate with several S-INAs.
  • a double shout auction embodies an interactive process between buyer and seller. In our example, a double shout auction can occur between a buyer and multiple sellers. Each negotiation process is two-way and allows multiple sessions. Ultimately, the multi-lateral approach will narrow the field as specific seller competitors drop out of the negotiation process after the selection of the winning seller.
  • Fig. 41 refers to the INA negotiation module and a negotiation method schema. The list of negotiation strategies and methods refers to which may influence the S-INAs first counter bid. This information input may involve collaboration so as to narrow the focus of customer data in order to facilitate customization. Given different information emphasis and the various item factors, each S-INA may provide quite different counter offer responses.
  • the B-INA (and its micro-agents) may provide second counter bids (3745 and 3750) to the S-INA counter offers. Again, the S-INAs respond with second counter offers (3755 and 3765). This process of counter- bidding and counter-offering may continue for numerous sessions, either with multiple sellers, two sellers or only one seller. The S-INAs may terminate the bidding/offering process at any time.
  • the B-INA continues to focus on the negotiation process with S-INA 2 by accepting the offer at 3775.
  • the contact is then sent to ITAs (3770 and 3780) for re-negotiation of some points or to final closure ofthe transaction, thus ending the negotiation session.
  • sequences can occur interactively between only one B-INA and one S-INA or between on B-INA and multiple S-INAs. This complex, multi-lateral, interactive negotiation process creates very dynamic scenarios, like occurring in one or multiple sessions.
  • Automated negotiation is illustrated in Fig. 38 in a demand-initiated sales system.
  • the sequences specified alternate between buyer and seller in a compromise process within pre-established parameters between one buyer (B-INA) and one seller (S-INA).
  • B-INA buyer
  • S-INA seller
  • the first part of the process negotiates a first variable
  • the second part negotiates a specific approaches and techniques that INAs may employ to automate negotiations in specific situations.
  • negotiations between a B- INA and an S-INA may be cooperative or competitive. If cooperative, the negotiation sessions can use either an exchange based approach or a problem-solving approach, as described. If competitive, the negotiation sessions are dialectical or oppositional, or deterrence based. Unlike other approaches, a deterrence negotiation approach uses a non-zero- sum game.
  • Negotiations may also be buyer-initiated or seller-initiated.
  • multi-lateral negotiations either one buyer can negotiate with multiple sellers, or several buyers can negotiate with multiple sellers. Whether one-to-one or multi-lateral, negotiations can cover different terms and goals of each party, as well as multiple item packages.
  • Automated negotiation can occur by establishing pre-set expert- system strategies in a game theoretic environment with specified constraints (i.e., time, information or choice).
  • constraints i.e., time, information or choice.
  • automated negotiations can be adaptive to constantly changing conditions. Such adaptation involves the anticipation of opponent potential activities, as well as of changing situations.
  • the addition of mobility creates another layer of complication for automated negotiation because location changes add logistical and sequential issues in the mechanics of negotiation operations.
  • Fig. 42 refers to the INA auction module, which specifies auction types. These auction categories may be employed by INAs in negotiation sessions.
  • the auction types can be mutually selected by buyer and seller(s) during the pre-negotiation stage of the negotiation process. Though they appear to be generally biased towards the seller side, these main auction types are all interactive, and may be used in conjunction with complementary auction types. For instance, an ascending auction when combined with a descending auction in an interactive environment, leads to a double shout auction.
  • a Vickrey auction merely modifies an English auction.
  • a Vickrey auction may alternate with an English auction as part of the overall "discount" method proposed by a seller to give it an advantage.
  • the negotiation module relies on information and analysis from the auction module and, in the case of S-INAs, the pricing module. Fig. 43 illustrates these interactive relations between the modules. The negotiation module also accesses Al when necessary. Once negotiation methods, strategies and tactics are selected by INAs, the INAs proceed to INA sessions (4112, 4117 and 4125). via the AAs. Both B-AAs and S-AAs interact with the negotiation module.
  • the pricing strategies module is shown in relation to AA interactions.
  • Market data (4155) informs competitor prices
  • INAs whether seller or buyer, do not need to have purely neutral stances from which to act.
  • INAs may have personalities or attitudes.
  • Fig. 45 refers to examples of personality traits that INAs may have as well as the dynamics of INA interaction.
  • INAs may be optimistic, opportunistic, and aggressive if sellers have shortages, or if buyers have surpluses.
  • INA personalities or attitudes may vary and alternate between the optimistic and the pessimistic, or between the opportunistic and the un-aggressive.
  • ITA functions can be included in an INA for concurrent program execution.
  • AA functions may be included in an INA.
  • These embodiments may include abbreviated versions of agents for enhanced efficiency of program code operation.
  • INAs use intelligence. The use and implications of applying Al to INAs provides an important layer of mobility which represents an additional embodiment.
  • C-INAs and Aggregation Methods Referring to Fig. 46, cooperative INA (C-INA) (4530) intermediation allows a neutral agency capacity by brokering negotiations between B-C- INAs and S-INAs.
  • a C-INA (4530) after accessing a CSA (4520) (which accesses various showcases and presents a report of an initial query), acts as a broker between S-INAs (4525) and multiple buyer C-INAs (4535). Because the B-C-INAs congregate for the pooling process and because they cooperate for the purposes of aggregating for better discounts, specific items from S-INAs may fill specific (multi-item) baskets of B-C-INAs at specified intervals by using ITAs (4540).
  • B-C-INAs can be essentially B-INAs that aggregate or work together in order to cooperate for discounts and more substantial buying power than individual B-INAs, there are several sources that initiate B-C- INA transactions. Fig. 47 describes these B-C-INA transaction initiation sources. In all cases, promotions from sellers are provided to, or invited by, sellers. However, in one embodiment, a B-C-INA may identify an opportunity that may require group buying power and inform other B-C-
  • INAs so as to pool a cooperative group for this opportunity.
  • Fig. 48 illustrates B-C-INA aggregation.
  • a CSA 4625
  • the lead B-C-INA (4635) interacts with other B-C-INAs so as to coordinate and prioritize their preferences.
  • At least two S-INAs (4850 and 4655) interact with B-C-INAs.
  • the lead B-C-INA may act as a sort of consolidator in this context, in effect providing initiation and clearinghouse agency functions.
  • Buyer lAs and pre-negotiation stages are used here in the preferred embodiment similar to ordinary B-INA and S-INA interaction early stage negotiation processes. As demonstrated in Fig.
  • B-C-INAs can initiate coordination when they realize common interests and communicate with each other.
  • this form of simple aggregation is merely a method to sell a quantity of items to multiple buyers.
  • B-C-INAs may be simultaneously coordinated for group buying opportunities. Any B-B-INA can broadcast opportunities to other B-C-INAs with similar interests. These broadcasts are sent to B-C-INAs through registries that identify and inform similarly interested parties.
  • the B-C-INA that broadcasts a buying opportunity then leads the aggregation process for its follower B-C-INAs.
  • B-C-INAs use tags to track their movements in the congregation process that precedes aggregation.
  • a competition between at least two S-INAs over comparable items ensures a competitive environment which provides greater value to B-C-INAs.
  • a seller may trigger buyer cooperation by initiating a promotion on items or packages focused on groups of buyers.
  • B-C-INAs may, in the course of negotiation with S-INAs, compromise in order to agree to the simplest items on their agenda, eventually filtering out the less common, mutually interested items. In this way, agreement between multiple buyers may be more easily and quickly reached.
  • ITAs whether buyer or seller - may require the completion of more tasks, in order to close the transaction (4675).
  • Aggregation in general, is a method to sell items, or bundles of items, to multiple buyers. Automated aggregation, in the context ofthe present system, may have several forms, including those shown in Fig. 49. The bundles can be multiple quantities of identical (or near identical) items as well as pre-set and open bundles. Pre-set bundles are specified combinations, while open bundles are any combination of items.
  • the distinctive distribution patterns of the various types of aggregation - listed in 4740 thru 4765 - each refers to a unique approach.
  • Figure 50 refers to the mass pooling method of automated aggregation using INAs with multiple buyers and multiple sellers.
  • INAs (4805, 4810, 4815, and 4820) pool their common interests by working together to procure specific items or bundles of items within predetermined time constraints (4825). As time deadlines pass, specific item sets are distributed from sellers to interested buyers (4830); such common bids and negotiations between buyers for seller items are distributed by sellers at various locations (4835). B-ITAs and S-ITAs process the orders or require further negotiation (4840 and 4865). Once approved by both sets of ITAs, final orders are distributed to buyers (4870) and the session(s) are closed after a quorum of items bought has been satisfied. In this way, multiple sellers sell to multiple buyers once specific constraints have been satisfied over time.
  • Fig. 51 illustrates the disintermediated method of aggregation employed by the present system prior to the negotiation phases.
  • several seller INAs 5005, 5010, and 5015
  • sell specific items 1-6
  • preset bundles i.e., specific combinations of items
  • S-INAs use S-AAs (5017) that use forecasting analysis of item combinations (5016) and demand shaping of time-sensitive promotional invitations (5020) (e.g., if a surplus of items creates an incentive by sellers to shift buyer demand from scarce items).
  • S-AAs inform promotions (5025) which are provided to the lead B-C-INA (5030).
  • the lead B-C-INA selects items (5035) that follower B-C-INAs (5045) may be interested in.
  • the B-C-INAs then enter into negotiations with the S-INAs for specific items and bundles (5055).
  • FIG. 52 refers to the aggregation process that provides disintermediated mass customization.
  • S-INAs (5105, 5110, 5115, and 5120) have items 1-8 and pre-set bundles 1-4.
  • B-C-INA 1 receives only item 1, but with features 2 & 3;
  • B-C-INA 2 receives item 1 with features 1 & 4 as well as item 4 with features 2 & 3;
  • B-C-INA 3 receives item 1 with features 1 & 4, item 4 with features 2 & 3 and bundle 3, and;
  • INA negotiation process is employed as well as the main aggregation process using B-C-INAs.
  • D-INAs D-INAs
  • D-INA dynamic INA
  • the D-INA switches roles (5265) and shifts to its seller mode (5275) moving to negotiate with a B-INA along with at least one other S-INA.
  • the buyer INA selects an item from a D-INA (now a seller) and after the ITAs resolve the closing of the transaction, in this illustration, the item can be provided to a buyer directly from the original seller (5295) thereby decreasing supply chain friction and duplication.
  • the transaction is then closed (5300).
  • the use of arbitrage involves the exploitation of limited buyer information from D-INA intermediaries.
  • One advantage of arbitrage approaches is the use of information at one location to exploit at a different location.
  • the present system - which uses mobile D-INAs in one embodiment - is particularly well suited to arbitrage approaches in geographically transcendent environments using D-INAs in their buyer and seller modes.
  • D-INAs use RMOs so as to limit risk. Precisely because there are enhanced risks in arbitrage situations between a seller and buyer function, RMOs in this context are compounded.
  • Fig. 63 refers to a method for disintermediated arbitrage of multi- item bundles from multiple sellers to multiple buyers using D-INAs.
  • S-C-INAs (6605, 6610, 6615, and 6620) provide multiple items for sale and cooperate in order to calculate multiple buyer bundle bids.
  • At least two D-INAs (6630 and 6635), after interacting with B-AAs (6625 and 6640), respectively request an initial search for multiple items via CSAs (6642 and 6414), and then enter into negotiations with the S-C-INAs (6630 and 6650).
  • Bids are evaluated at 6655 using either relationship management (6660) or revenue maximization (6665) strategies, after which the D-INAs select sellers' specific bundles (6670). Once a seller bundle or combination of seller items are selected by D-INAs (as buyers), the D-INAs change their mode to that of a seller
  • the D-INAs then negotiates (as a seller) with multiple B-INAs for multiple items. In this example, several B-INAs are winnowed in succeeding phases until the final B-INAs are selected.
  • the application of D-INAs for multi-items in the seller mode reveals a dis-aggregation function by selling to several buyers.
  • a traditional, intermediated, combinatorial auction is illustrated with an application to a single seller providing items to multiple sellers.
  • a seller (5505) provides multiple items, 1-5 (5510-5530), via an intermediary (5535), to various B-INAs.
  • the items are distributed in this example, according to specific combinations of items, to specific B-INAs.
  • B-INA 1 receives items 2 & 4
  • B-INA 2 receives items 1, 2 & 3, etc.
  • the intermediary is removed from the transaction in which a seller provides four items to four separate B-INAs in specific configurations.
  • Buyer A receives items 1 & 3, buyer B items 1 & 2, buyer C items 2 & 4, and buyer D items 2, 3, and 4. Because no intermediary is involved, a double opposing shout auction - in which the package price descends for seller(s) while it simultaneously rises for the buyer - is used between a single seller and several buyers simultaneously.
  • Fig. 56 illustrates a multilateral opposing shout auction in which items are sold between a buyer and at least two sellers, either with or without an intermediary. From the viewpoint of a seller, prices decline, while from the viewpoint of the buyer, prices increase.
  • Factor filters are methods by which to prioritize multi-item bundles by composition and structure. By distinguishing between kinds of bundles, negotiation for multiple items between buyer(s) and seller(s) can be more organized and efficient. Such factor filtering processes can be applied to combinatorial auctions employing INAs.
  • factor filters operate as pruning techniques (5815) in the process of evaluating multiple bidders (5810) by either S- INAs or B-INAs (5805). After establishing a priority preference (5820), several main kinds of bundles - preset (5825), specific (5830), progressive (5835), quantity (5840), quality (5845), and temporal adaptive (5850) - are categorized.
  • threshold factor specific i.e., an item that is critical to a bundle
  • need specific (5900), in which complementary item(s) are necessary in order to make the whole bundle desirable
  • item preference (5890), in which a preferred item in a bundle is sought
  • a D-INA may buy a bundle for a specific item in order to split the bundle and resell the various valuable and common parts.
  • successor contingent bundles (5860) in which a first priority item is sought and only if not acquired then a second priority item is sought and so on
  • priority contingent bundles (5865), in which a first item is sought and, only if the first item is acquired, will a second item be acquired and so on
  • quantity bundles (5895) in which multiple substitutable items are acquired, for example by more than one buyer as an aggregate
  • quality bundles in which the best items are sought (5867) and for the best value (5883)
  • dynamic pricing contingent bundles in which multiple items
  • Fig. 58 disintermediated multi-item bidding from only one seller to multiple sellers is described.
  • the B-INAs After several B-INAs receive data analysis from B-AAs, the B-INAs enter into mutually agreed rules of negotiation (5995) similar to a pre-negotiation.
  • a seller INA negotiates with the B-INAs for multiple items (6000) by proceeding to apply factor filters (6005) in order to establish buyers' specific priorities.
  • Buyers (6007) select unique sets of specific items in order for the seller to evaluate the initial bids (6015) for optimal seller benefit.
  • the seller uses two main approaches. On the one hand, it can use a strategy of short-term revenue maximization (6020) in which it accepts the overall two highest bids (6040) for a specific package. On the other hand, it can use a strategy of long-term relationship management (6010) in which two winners are selected by using factors beyond price alone. In an additional embodiment, it can select two winners by alternating between the two methods. Once the two winning buyers are determined (6030 and 6035) for multiple items, the process enters a new phase. A second phase of bids are evaluated (6045) and the highest overall bid on multiple items (6050) by a buyer is evaluated by the seller.
  • a winner is determined by the S- INA (6055) and the ITAs close the transaction (6057).
  • the second highest bidder can capture remaining seller items not included in the first winner package and hence constitute another package of items. The second highest bidder then becomes the winner of a second priority bundle of items remaining from the first bundle of buyers.
  • Fig. 59 describes a method for transactions involving multi-item bidding with multiple sellers to a single buyer using S-INAs.
  • a B-INA (6110) requests bids from sellers for specific bundles of items.
  • S-INAs work together (i.e., cooperate to mutually agree on negotiation rules (6115) similar to a pre- negotiation session.).
  • bidding occurs by the S-C-INAs to supply packages of items from various sellers.
  • Pre-set bundles (6150) are bid on and a winner determined by the buyer (6160).
  • Pruning techniques (6155) that eliminate less preferred items and factor filters (6170) are applied in order to limit bundle composition so as to increase efficiency.
  • seller bids are evaluated by the buyer.
  • the buyer can use a short-term revenue maximization (6180) strategy or a long-term relationship management (6165) strategy of preliminary winner determination.
  • the former strategy leads to S-INA 4 being selected and the latter strategy leads to S-INA 2 being selected.
  • a final winner - S-INA 2 (6210) - is selected.
  • a B-ITA resolves any transaction clearing issues (6205) and either renegotiates or closes the transaction (6200).
  • Fig. 60 shows a process for disintermediated aggregation of pre- set bundles with multiple sellers and multiple buyers.
  • Pre-set bundle 1 (6260) is provided by seller 1 (6255)
  • pre-set bundle 2 (6275) is provided by seller 2 (6270)
  • pre-set bundle 3 (6290) is provided by seller 3 (6285).
  • Pre-set bundle one consists of products 1 & 2 and service 1
  • pre-set bundle two consists of products 3 & 4 and service 2
  • pre-set bundle three consists of products 5 & 6 and service three.
  • buyer one (B-C-INA 1) and buyer two (B-C-INA 2) select pre-set bundle one from seller one.
  • Buyers one, two, and three also select pre-set bundle two from seller two. All buyers select pre-set bundle three.
  • Fig. 61 shows a disintermediated bidding approach for multiple items between multiple sellers and multiple buyers.
  • sets of combinations of items are matched between sellers and buyers.
  • Each horizontal row represents the distinct items offered from one seller.
  • row one represents items 1-4 from seller one and so on.
  • five buyers bid for separate specific packages (bundles of items) from among the twelve items offered from the three sellers. Accordingly: Buyer Bidder A seeks items 1 , 5, & 9 Buyer Bidder B seeks items 2, 3, 4, 6, 7, & 8
  • Buyer Bidder C seeks items 3, 4, 7, 8, & 12
  • Buyer Bidder D seeks items 5, 6, 7, 9, 10, & 11
  • Buyer Bidder E seeks items 9, 10, 11 , & 12
  • Fig. 62A illustrates C-INAs used on both the buyer side and the seller side.
  • a disintermediated method of aggregation is described involving multiple item bidding from multiple sellers and multiple buyers using C-INAs.
  • various sellers 6415, 6435, 6455, and 6470
  • B-C-INA 1 (6430) is the lead C-INA in this example.
  • B-C-INAs congregate at various locations at the request ofthe lead B-C-INA.
  • a pre-negotiation phase will set initial rules.
  • Each B-C-INA seeks different sets of items from various sellers.
  • the sellers In order to bid on a variety of items requested from multiple buyers offered by multiple sellers, the sellers must work together. At 6445, the sellers cooperate by providing pricing information in order to calculate B-C-INA bid values. Without this cooperation, incomplete information on items in bid sets not involving only a specific seller will be indeterminable. Though complex, the goal of sharing pricing information between sellers regarding buyer bidding is to manage auction pricing (and other item factor) data in a limited time frame so as to establish a competitive real-time market.
  • INAs that pool their interests (6510) into multiple item bundles.
  • Bids are made and evaluated (6530) by buyers either cooperatively or competitively; if cooperatively, buyers may divide items between themselves after sellers have determined the buyer winners.
  • the group of buyers may also assemble specific subsets of items comprising specific packages (6455). In this case the buyers may bid for a specific subsets of items (6575) for distribution from among participants after the general sale from a seller. This distribution process is a form of dis-aggregation.
  • the selection of successful B-C-INAs may occur over time by instituting overlapping time frames for the filling of buyer baskets.
  • Fig. 64 refers to various factors that may be negotiated in multi- item negotiations.
  • Fig. 65 refers to a list of alternative multiple feature factors involving personal computer configurations. A change in a single feature changes the composition of the package. Each completed computer represents a multi-item package. If different sellers provide the pieces of each computer, then a buyer using the present system can assemble a multi-item bundle.
  • Fig. 66 illustrates examples of multi-item bundle categories.
  • Intelligent transaction agents are used to close transactions.
  • ITAs are either seller-side or buyer-side.
  • ITAs (6815) interact with AAs (6810) and with INAs (6805) as illustrated in figure 67 in the context of the seller role.
  • the ITA closes the transaction (6830) after it clears the negotiation.
  • the ITA uses a compliance function, in these cases.
  • Fig. 68 shows an S-ITA operation. After an S-INA provisionally completes negotiation (6850), it requests the S-ITA to clear the transaction (6855). The S-ITA initiates a review of transaction terms and item terms (6850), and accesses a financial database to check the buyers credit (6865). The S-ITA proceeds to clear the buyers credit (6870 and
  • the INA and ITA work together symbiotically to clear each variable in a sequence of transaction steps. As the INA requires an ITA function, it will pass this part of the transaction for clearance while the INA continues to process the negotiation functions contingent on ITA clearance.
  • the ITA represents an autonomous sequential clearing mechanism in a close relationship with the INA operation. ITAs provide an important function similar to an accountant or lending officer. Without clarifying item and transaction terms, for instance, an agreement is not complete. Once clarified by using a checklist of operations that pertain to important functions, the deal can be closed.
  • Fig. 69 describes an S-ITA (6940) and a B-ITA (6945) interacting in a system process in the final negotiation between one seller and one buyer.
  • the ITAs interact with their respective AAs (6920 and 6950) and INAs (6925 and 6930) and with each other. If the respective INA does not satisfy ITA transaction clause constraints, the transaction is returned to the INA for renegotiation of its specific parts during a specific phase of buyer-seller interaction. Once all constraints are satisfied, the transaction is closed (6955).
  • ITAs may perform their functions at multiple locations, in sequence or alternating at various locations. This mobility aspect is achieved by using Al technologies. ITAs also supply services as listed in Figs. 70 and 71. At each stage of a sequence of INA interactions, the ITA may offer these services to buyers and sellers.
  • ITAs can analyze data crucial for transaction completion. Consequently, ITAs can involve services referenced in figs. 88-92.
  • ITA functions may be included in INAs for optimized efficiency and may be executed concurrently.
  • a B-INA negotiates with at least two S-INAs to conduct multiple parallel (or sequential) negotiations.
  • One method to do this is for a B-INA to spin off various "children" or B- INA micro-agents (7110, 7115, and 7120).
  • Each B-INA micro-agent can complete a negotiation session with one of a series of S-INAs at various locations (7125, 7130 and 7135).
  • Micro-agents can communicate with each other in order to conduct parallel simultaneous negotiations with multiple S-INAs.
  • Micro-agents may use applets or aglets in order to launch, replicate and activate their program code.
  • D-INAs particularly in a buyer mode, may use micro-agents to conduct its functions.
  • S-INAs may use micro-agents to conduct its functions.
  • micro-agents may be applied in a similar way.
  • Micro-agents are particularly appropriate in complex multi-lateral negotiation activities in which mobility of agents between multiple locations in a distributed network are involved.
  • Figs. 39A and 39B illustrate a mobile negotiation method in a distributed system that can be applied to micro-agents. Artificial Intelligence
  • Al Artificial intelligence
  • AAs INAs
  • ITAs ITAs
  • CSAs CSAs
  • Figs. 73-80 describe the unique operations of Al and their distinctive applications to agents in the present system.
  • agent actions can be autonomous.
  • Such autonomous agency provides unique interactions among agent operators that emulates the complexity of markets.
  • the present system advances considerably the use of Al in multi-agent computer program commercial systems.
  • a search for the best algorithmic solution commences (7215). Based on available information, a set of candidate solutions is generated (7225). In a distributed communications system, market data inputs (7210) are filtered (7220). Candidate solutions (7225) are created from available information and then new generations of candidate solutions based on multi-factorial mutations (7235) that are generated by a randomized mutation engine (7228). Mutated candidate variants are produced using survivor candidate operators that sort by variables (7240). These maximized and expanded sets of candidate solutions are evaluated according to efficiency criteria (7245), which can be used to select the most optimized candidate solutions. The best algorithmic candidates are matched with market data (7230) [via pattern matching (7220)] and then new generations of candidate solutions are created in a loop from 7230 to 7245.
  • the winning algorithm solution is determined using the most updated criterion (7260).
  • the solution results are displayed (7270) and ranked (7280).
  • Sub-optimal solutions (within specified constraints) are returned to generate additional candidate variants for future comparison, selection and use.
  • the optimal solution, relative to all available candidates (7275), is applied to agent analysis or activity (7285) and the program run is ended (7290).
  • NNs are generated (7325), in part based on available market data and in part based on a comparison of optimal statistical scenarios (7330). Statistical scenario comparisons may involve fuzzy logic inputs (7320).
  • Neural networks are trained (7335) for fitness using training patterns that run through a process of trial and error until a specific set of candidate NNs is identified that optimize computation solutions. These patterns are compared to market data inputs. NNs are matched for optimal fitness patterns (7340). During this process of pattern matching, mutations may occur using a mutation engine (7345) that employs random (7355) and alternating sequences (7360). Such mutated NNs are retrained using efficient parallel computation resources. The most fit NNs (7340) are pruned (7350), tested (7365) and ranked.
  • Fig. 75 a genetic programming system process is described.
  • Data inputs (7505) are applied to rule based learning (7510), regression analysis (7515) and induction decision trees (7520).
  • Rule based learning uses a pattern matching pruning approach (7525) that leads to the development of heuristic operational rules (7535) that relearn (7545).
  • the heuristic operational rules, the regression analysis, and the induction decision trees are applied to organizing models (7530).
  • the organizing models are measured by scope, accuracy, and errors (and exceptions and missing values).
  • These operating models present output scenarios (7550) as statistical positive (7560), or negative (7565) recommendations or as forecasts (7555). These recommendations are applied to agent analysis or activity (7570) and the computer program run is ended (7575).
  • GP learning schemas are described in Fig. 76 as a table of various main learning types.
  • Fig. 77 describes evolutionary computation applications to agents in a multi-agent system. After testing GA approaches for success, the system proceeds to test GP, and NN approaches. This filtering approach of testing EC techniques operates like a switch.
  • the evolutionary computation approaches (7605) of genetic algorithms (7610), genetic programming (7615) and neural networks (7620) are applied to the CSA (7625), the AA (7630), the INA (7635), and the ITA (7637).
  • Each agent uses differentiated reasoning schemas that are specified.
  • Each agent type also specifies the advantages of each application (7640, 7645, 7650, and 7655).
  • Fig. 78 describes Al applied to agents in a distributed system.
  • An agent requests (7675) and accesses (7680) Al application to solve a problem at a particular location.
  • the agent then identifies optimal Al application by using a filtering process (illustrated in Fig. 77) that test GA, GP and NN processes for a specific use (7685).
  • the agent then applies Al efficiently for a specific use (7695). If an agent requires more Al
  • an evolutionary computation architecture is described with reference to AA and INA applications.
  • GA (7715), NN (7725), and GP (7720) modules breed optimal programs (7730) using the efficiency module (7735).
  • Al is applied in distinctive ways to the present system. Techniques empower agents to be autonomous and, hence, mobile. EC endowed autonomous intelligent software agency is applicable to negotiation agents using several specific methods.
  • Al is applied to: (1) the B-INA process of narrowing from several S-INAs to two S-INAs; (2) the process of a B-INA automatically negotiating with two S-INAs, (3) the process of interactive multivariate B- INA and S-INA negotiation and (4) the process of using mobility in automated demand-initiated negotiations in a distributed environment.
  • Al-induced INAs can solve complex negotiation problems within specific rules that pre-programmed systems cannot. Since the present invention involves several dimensions of complexity, including multivariate, multilateral, combinatorial and mobile aspects in a demand-initiated system - which are exponentially more complex in combination - Al approaches are increasingly pivotal. Mobility
  • Fig. 80 refers to layered Al for optimal agent mobility.
  • the system resembles a RISC (reduced instruction set computing) software architecture approach that strongly benefits the need to keep mobile agent program code as efficient as possible.
  • huge computer resources are accessible when necessary - either from the users "home" computer or an outside service provider - in order to provide powerful computation to meet peak agent needs, especially in time-sensitive or complex transactions.
  • This approach may resemble a biological immune system which, when it detects an anomaly, brings to bear a larger arsenal.
  • the application of layered Al promotes the use of mobile program code so as to efficiently enhance scarce resources.
  • intelligent agents have a "home" base (7860), or a computer source location from which it is launched (7870). After launching, the agent(s) make an initial determination of minimum expected computer resources required for a specific activity based on initial (pre-negotiation) interaction (7885). After the initial determination of efficient computer resources, the agent(s) enter into interactions (or analysis) using specified Al function levels at various mobile locations (7895) specified in the transaction(s). Either the agent(s) completes their function(s) and terminates (7897), or they discover a need for increased Al to perform increased functions (7900).
  • the agent(s) seek "reserve" layers of Al (7880). These reserve layer requests are analyzed for minimum actions necessary (7865) to perform a specific function efficiently (thereby accommodating mobility requirements), parallel use (7875) or alternating use (7890) of various Al simultaneously for faster and effective computation resource capacity utilization. These added resource capacities are input at the agent interaction level (7895).
  • Al filters are employed in which the agent(s) select optimal methods of Al to employ (7905).
  • an INA may use GP and NN computing preferences to complete negotiations, while an AA or ITA may prefer to employ GA to enhance an analytical function.
  • the optimal Al method request moves to an agent requesting reserve Al layers (7880).
  • the agent(s) seek out the most effective use of Al functions (7915).
  • an agent needs to determine a (constantly shifting) balance; either (7920) between less Al program code in order to maintain high speed and light travel, or increased Al sophistication for intensive analysis or negotiation activities.
  • the agent disables unnecessary code (7930).
  • the buyer interacts with a B-AA (8035) and a CSA (8060) directly, and via a B-IA (8015).
  • the B-AA interacts with the CSA (8060), B-INAs (8070) and B-ITAs (8065).
  • AAs may be mobile. Both S-AAs and B-AAs use Al, which may involve a need to return to home base (customer or seller) for increased computation resources.
  • Collaborative filtering uses statistical scenarios (8145), forecasting (8185) and syntheses (8150) methods, which result in issuing a recommendation report (8200).
  • Editorial retailing involves ascertaining third party opinions (8155) and then initial filtering and synthesis of buyer data (8170), the combination of the data (8190) and production of a final report (8205).
  • the reports - typically customized - are made available to B-MNAs and B-C-MNAs.
  • the expert system involves applying pre-programmed parameters to the mass personalization of data (8160), the use of targeted information (8175) and the creation of a custom report (8195).
  • Multi-attribute regression analysis typically isolates variables according to an established relevance scale by using a process of "factor testing” that measures the accuracy of specific attributes. At a stable equilibrium point in the analysis, a synthesis of attributes (8165) can be made that combines key variables for maximum utility, after which a systematic report is generated (8180).
  • Fig. 83 refers to the analytical agent data flow process.
  • Market data (8255) from various sources in a distributed computer system, is translated into codes.
  • AAs (8260) access the market data and provide a range of services (8265) specified at 8270 and 8275 as well as figures 92.
  • Fig. 84 data mining approaches are described with particular reference to CSA and AA interactions.
  • Market data (8305) is fed into the system from different sources, while various methods of data discovery are employed (8310).
  • these methods include: search (8315) - including both local (8340) and global (8375) databases - specific (8320) query (by item (8345) and by company (8350) accessible from showcases (8380)), general (8325) query (by category (8355) and industry (8360)), time sensitive query (8330) and targeted information collection (8335) (that may involve specific customer requirements (8370)).
  • AAs use data syntheses (8385) that create specific customer profiles (8400), and data analysis (8390) that features predictions (8395) and creates scenarios (8410).
  • a report is created for AA use.
  • a process for advanced collaborative filtering for cross-marketing recommendations is described in Fig. 85.
  • the customer requests information on item(s) of interest (8470).
  • the collaborative filtering process sorts categories according to item type, popularity, region, quality, services, bundles, quantity, price range, and combinations of these categories.
  • the customer request for information is analyzed (8500) and new items of customer interest are statistically ranked in relation to the initial item request. Other items are recommended that are associated with information on the current item (8510). A list of recommended items is presented to the customer (8515) who subsequently acquires them
  • Fig. 86 shows B-AA operations with mobility.
  • a B-INA or B- ITA
  • B-AA requests analysis from a B-AA (8550)
  • the B-AA activates analysis functions at a specific location (8555).
  • the B-AA then moves to a remote or multiple remote locations to collect data (or it may import data at its home location) (8560).
  • the B-AA performs analysis on data (8570), organizes the analysis and issues a report (8575).
  • the B-AA also performs synthesis on the data by combining data sets (8580) and organizing the data synthesis into a report (8585).
  • the data analysis and synthesis functions are then applied by the B-AA to a B-INA (or BITA) function at a specific remote location by exporting reports (8590) and then closes the session (8600).
  • AA functions may be consolidated with or included in INAs or ITAs and may execute concurrently.
  • Figures 87 through 91 describe a system involving services and service variables that utilize a distinctive code process. Most agent interaction involves the exchange of information using these codes.
  • Fig. 92 lists services provided in the present system.
  • the system represented by the present invention has numerous distinctive embodiments.
  • the present disclosures illustrate in detail the main ideas of the invention and are not intended to restrict the invention to a single embodiment.

Landscapes

  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Computer And Data Communications (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

L'invention concerne des agents intelligents de négociation (INA) (1085,1090,1080) situés dans un système réparti à multi-agents ouvert à la demande, dans lequel un agent acheteur négocie avec au moins deux agents vendeurs. Ces agents intelligents de négociation sont des agents intelligents autonomes de logiciel qui adaptent les paramètres de négociation pour l'acquisition de produits, de services et des groupes d'articles divers en appliquant des technologies d'intelligence artificielle. Des services, notamment des promotions de marketing, des options de gestion de risques (1020) et des services de fait sur commande sont intégrés au système. L'agrégation accessoire de l'acheteur se produit par mise en commun des agents intelligents de négociation de coopérative (c-INA). Des agents intelligents de négociation dynamiques agissent en tant qu'agents doubles qui alternent les rôles d'acheteur et de vendeur, ce qui permet l'arbitrage. Des processus de collaboration (1035) facilitent des interrogations multilatérales personnalisées. Cette interaction de données précède des négociations automatiques (1065, 1070). L'invention concerne un réseau réparti spécifique à une industrie de bases de données de présentoir de coopérative dans lequel les données sont constamment mises à jour, analysées et nettoyées par des inter-agents de présentoir (1015, 1050, 1040). Des agents de recherche commerciale (1060) (CSA) font des demandes d'articles pour une première demande d'un vendeur dans un réseau de communication commercial (CNN). Des agents analytiques (AA) (1010, 1013, 1095, 1100) déposent des données. Les agents peuvent être mobiles.
PCT/US2001/047023 2000-12-01 2001-12-03 Agents intelligents de negotiation ouverts a des demandes situes dans un systeme reparti WO2003067494A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
AU2001297749A AU2001297749A1 (en) 2000-12-01 2001-12-03 Demand-initiated intelligent negotiation agents in a distributed system
GB0314327A GB2390194A (en) 2000-12-01 2001-12-03 Demand-initiated intelligent negotiation agents in a distributed system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US25081900P 2000-12-01 2000-12-01
US60/250,819 2000-12-01

Publications (1)

Publication Number Publication Date
WO2003067494A1 true WO2003067494A1 (fr) 2003-08-14

Family

ID=27662793

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2001/047023 WO2003067494A1 (fr) 2000-12-01 2001-12-03 Agents intelligents de negotiation ouverts a des demandes situes dans un systeme reparti

Country Status (3)

Country Link
AU (1) AU2001297749A1 (fr)
GB (1) GB2390194A (fr)
WO (1) WO2003067494A1 (fr)

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7467196B2 (en) 2005-01-12 2008-12-16 International Business Machines Corporation Managing network errors communicated in a message transaction with error information using a troubleshooting agent
US7472079B2 (en) 2005-01-12 2008-12-30 International Business Machines Corporation Computer implemented method for automatically controlling selection of a grid provider for a grid job
US7562035B2 (en) 2005-01-12 2009-07-14 International Business Machines Corporation Automating responses by grid providers to bid requests indicating criteria for a grid job
US7571120B2 (en) 2005-01-12 2009-08-04 International Business Machines Corporation Computer implemented method for estimating future grid job costs by classifying grid jobs and storing results of processing grid job microcosms
US8136118B2 (en) 2004-01-14 2012-03-13 International Business Machines Corporation Maintaining application operations within a suboptimal grid environment
US8275881B2 (en) 2004-01-13 2012-09-25 International Business Machines Corporation Managing escalating resource needs within a grid environment
US8387058B2 (en) 2004-01-13 2013-02-26 International Business Machines Corporation Minimizing complex decisions to allocate additional resources to a job submitted to a grid environment
US8583650B2 (en) 2005-01-06 2013-11-12 International Business Machines Corporation Automated management of software images for efficient resource node building within a grid environment
WO2013192314A1 (fr) * 2012-06-19 2013-12-27 Visible World, Inc. Systèmes, procédés et supports lisibles par ordinateur pour optimiser des transactions dans un réseau multimédia adressable domestique
WO2016134761A1 (fr) 2015-02-25 2016-09-01 Siemens Aktiengesellschaft Procédé de fabrication d'un produit selon un plan de production comprenant une pluralité d'étapes de production, et un installation de fabrication
RU2662638C2 (ru) * 2016-12-29 2018-07-26 Валерий КАРДАНОВ Способ одновременного запроса скидки у нескольких поставщиков на электронной площадке агрегатора товаров
WO2019088872A1 (fr) * 2017-10-31 2019-05-09 Андрей Александрович ТУНИКОВ Procédé et système pour effectuer des enchères électroniques en ligne
WO2019088871A1 (fr) * 2017-11-02 2019-05-09 Андрей Александрович ТУНИКОВ Procédé et système pour effectuer des enchères électroniques en ligne
US10803535B2 (en) 2017-04-20 2020-10-13 International Business Machines Corporation Facilitating power transactions
US11328228B2 (en) 2019-04-22 2022-05-10 International Business Machines Corporation Location allocation planning

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2385954A (en) 2002-02-04 2003-09-03 Magenta Corp Ltd Managing a Virtual Environment

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1993021586A1 (fr) * 1992-04-15 1993-10-28 Inference Corporation Element de raisonnement et d'apprentissage autonome
WO1995011560A1 (fr) * 1993-10-21 1995-04-27 Martino John A Ii Systeme et technique d'interfacage de programmation d'application
US5421013A (en) * 1993-07-08 1995-05-30 Park City Group, Inc. Agent-based multithreading application programming interface
WO1995015635A1 (fr) * 1993-11-30 1995-06-08 British Telecommunications Public Limited Company Gestion de reseaux de communications
US5655081A (en) * 1995-03-08 1997-08-05 Bmc Software, Inc. System for monitoring and managing computer resources and applications across a distributed computing environment using an intelligent autonomous agent architecture
WO1998043146A2 (fr) * 1997-03-21 1998-10-01 International Business Machines Corporation Agent intelligent possedant des capacites de negociation et procede de negociation au moyen de cet agent intelligent
US6141653A (en) * 1998-11-16 2000-10-31 Tradeaccess Inc System for interative, multivariate negotiations over a network

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1993021586A1 (fr) * 1992-04-15 1993-10-28 Inference Corporation Element de raisonnement et d'apprentissage autonome
US5421013A (en) * 1993-07-08 1995-05-30 Park City Group, Inc. Agent-based multithreading application programming interface
WO1995011560A1 (fr) * 1993-10-21 1995-04-27 Martino John A Ii Systeme et technique d'interfacage de programmation d'application
WO1995015635A1 (fr) * 1993-11-30 1995-06-08 British Telecommunications Public Limited Company Gestion de reseaux de communications
US5655081A (en) * 1995-03-08 1997-08-05 Bmc Software, Inc. System for monitoring and managing computer resources and applications across a distributed computing environment using an intelligent autonomous agent architecture
WO1998043146A2 (fr) * 1997-03-21 1998-10-01 International Business Machines Corporation Agent intelligent possedant des capacites de negociation et procede de negociation au moyen de cet agent intelligent
US6141653A (en) * 1998-11-16 2000-10-31 Tradeaccess Inc System for interative, multivariate negotiations over a network

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8387058B2 (en) 2004-01-13 2013-02-26 International Business Machines Corporation Minimizing complex decisions to allocate additional resources to a job submitted to a grid environment
US8275881B2 (en) 2004-01-13 2012-09-25 International Business Machines Corporation Managing escalating resource needs within a grid environment
US8136118B2 (en) 2004-01-14 2012-03-13 International Business Machines Corporation Maintaining application operations within a suboptimal grid environment
US8583650B2 (en) 2005-01-06 2013-11-12 International Business Machines Corporation Automated management of software images for efficient resource node building within a grid environment
US7571120B2 (en) 2005-01-12 2009-08-04 International Business Machines Corporation Computer implemented method for estimating future grid job costs by classifying grid jobs and storing results of processing grid job microcosms
US7562035B2 (en) 2005-01-12 2009-07-14 International Business Machines Corporation Automating responses by grid providers to bid requests indicating criteria for a grid job
US8346591B2 (en) 2005-01-12 2013-01-01 International Business Machines Corporation Automating responses by grid providers to bid requests indicating criteria for a grid job
US7472079B2 (en) 2005-01-12 2008-12-30 International Business Machines Corporation Computer implemented method for automatically controlling selection of a grid provider for a grid job
US7467196B2 (en) 2005-01-12 2008-12-16 International Business Machines Corporation Managing network errors communicated in a message transaction with error information using a troubleshooting agent
WO2013192314A1 (fr) * 2012-06-19 2013-12-27 Visible World, Inc. Systèmes, procédés et supports lisibles par ordinateur pour optimiser des transactions dans un réseau multimédia adressable domestique
WO2016134761A1 (fr) 2015-02-25 2016-09-01 Siemens Aktiengesellschaft Procédé de fabrication d'un produit selon un plan de production comprenant une pluralité d'étapes de production, et un installation de fabrication
US10816961B2 (en) 2015-02-25 2020-10-27 Siemens Aktiengesellschaft Method for manufacturing a product according to a production plan
RU2662638C2 (ru) * 2016-12-29 2018-07-26 Валерий КАРДАНОВ Способ одновременного запроса скидки у нескольких поставщиков на электронной площадке агрегатора товаров
US10803535B2 (en) 2017-04-20 2020-10-13 International Business Machines Corporation Facilitating power transactions
WO2019088872A1 (fr) * 2017-10-31 2019-05-09 Андрей Александрович ТУНИКОВ Procédé et système pour effectuer des enchères électroniques en ligne
RU2730408C2 (ru) * 2017-10-31 2020-08-21 Андрей Александрович ТУНИКОВ Способ проведения электронных онлайн торгов на электронной торговой площадке и автоматизированная онлайн система для его осуществления
WO2019088871A1 (fr) * 2017-11-02 2019-05-09 Андрей Александрович ТУНИКОВ Procédé et système pour effectuer des enchères électroniques en ligne
US11328228B2 (en) 2019-04-22 2022-05-10 International Business Machines Corporation Location allocation planning

Also Published As

Publication number Publication date
GB2390194A (en) 2003-12-31
AU2001297749A1 (en) 2003-09-02
GB0314327D0 (en) 2003-07-23

Similar Documents

Publication Publication Date Title
US8478677B2 (en) System for intelligent search agent to access data in a distributed network
US20020069134A1 (en) System, method and apparatus for aggregation of cooperative intelligent agents for procurement in a distributed network
US20030233305A1 (en) System, method and apparatus for information collaboration between intelligent agents in a distributed network
US20030074301A1 (en) System, method, and apparatus for an intelligent search agent to access data in a distributed network
US20020055903A1 (en) System, method, and apparatus for a cooperative communications network
He et al. On agent-mediated electronic commerce
Chen et al. Bi‐negotiation integrated AHP in suppliers selection
Kalakota et al. Move over, e-commerce
US20110238521A1 (en) Method, System and Business Model for a Buyer's Auction with Near Perfect Information Using the Internet
Hawkins The phantom of the marketplace: searching for new e-commerce business models
WO2003067494A1 (fr) Agents intelligents de negotiation ouverts a des demandes situes dans un systeme reparti
Elmaghraby Auctions and pricing in e-marketplaces
WO2001033464A1 (fr) Systeme lance sur demande du client et procede pour la recherche d'informations en ligne, la negotiation interactive, l'acquisition, et l'echange
Klusch Agent‐Mediated Trading: Intelligent Agents and E‐Business
Agrawal et al. Matching intermediaries for information goods in the presence of direct search: an examination of switching costs and obsolescence of information
Pivk et al. Intelligent Agents in E-commerce
Argoneto et al. Innovative tools for business coalitions in B2B applications: How negotiation, auction and game theory can support small-and medium-sized business in E-business
Sun et al. Multiagent brokerage with CBR
Hilsen Simulating dynamic pricing algorithm performance in heterogeneous markets
Gerding Autonomous agents in bargaining games: an evolutionary investigation of fundamentals, strategies, and business applications
CA2389828A1 (fr) Appareil de negociation
Hershberger Exchanges for complex commodities: Toward a general-purpose system for on-line trading
GB2395327A (en) Method and system for online procurement
Choi et al. Product matching algorithm for cooperative commerce model
Chamodrakas et al. “Liquid” Electronic Marketplaces

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

ENP Entry into the national phase

Ref document number: 0314327

Country of ref document: GB

Kind code of ref document: A

Free format text: PCT FILING DATE = 20011203

Format of ref document f/p: F

121 Ep: the epo has been informed by wipo that ep was designated in this application
REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

122 Ep: pct application non-entry in european phase
NENP Non-entry into the national phase

Ref country code: JP

WWW Wipo information: withdrawn in national office

Country of ref document: JP