CN108910630B - Elevator group control system and method based on multi-agent competition mode - Google Patents
Elevator group control system and method based on multi-agent competition mode Download PDFInfo
- Publication number
- CN108910630B CN108910630B CN201810736500.XA CN201810736500A CN108910630B CN 108910630 B CN108910630 B CN 108910630B CN 201810736500 A CN201810736500 A CN 201810736500A CN 108910630 B CN108910630 B CN 108910630B
- Authority
- CN
- China
- Prior art keywords
- bidding
- information
- unit
- agent
- service
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B1/00—Control systems of elevators in general
- B66B1/02—Control systems without regulation, i.e. without retroactive action
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B1/00—Control systems of elevators in general
- B66B1/34—Details, e.g. call counting devices, data transmission from car to control system, devices giving information to the control system
- B66B1/3415—Control system configuration and the data transmission or communication within the control system
Abstract
The invention discloses an elevator group control system and method in a multi-agent competition mode, wherein each elevator control system is an autonomous agent, the agent obtaining a calling signal generates bidding information according to a preset protocol and sends the bidding information to other agents in a cooperative range, each agent bids aiming at the bidding information, and the agent with the optimal bidding price provides service for passengers; each agent includes: a user interface unit for interacting with a passenger; the communication unit is used for carrying out information transmission with other intelligent agents in the cooperative range; the bidding unit is used for generating all information in the bidding process and bidding with other intelligent agents in the cooperative range; and the control unit operates the plan and executes the service according to the call form after winning the bid. The invention adopts a non-centralized control mode independent of additional group control equipment, changes the unified dispatching of a plurality of elevators into the cooperative service of the plurality of elevators through a bidding mechanism, improves the reliability of the elevator group control system and reduces the requirements on software and hardware.
Description
Technical Field
The invention relates to a dispatching technology of multiple elevators, in particular to an elevator group control system and method based on a multi-agent competition mode.
Background
The elevator group control system enables a plurality of elevators in a dispatching range to operate cooperatively through a reasonable dispatching method, so that the overall service efficiency of the elevator group control system is improved, and more energy-saving and more comfortable dispatching requirements are generated in recent years.
A common implementation of a group control system for elevators is to add a group control unit with a dispatching function, where the group control unit communicates with each elevator in the dispatching area, each elevator sends a service request and its own status to the group control unit, and the group control unit sends a group control dispatching command to each elevator, which is collectively referred to as a centralized control mode. In the actual operation process, once the group control unit fails, the whole group control system cannot operate, and the elevator dispatching is paralyzed, which is a fatal problem of the elevator group control system in a centralized control mode.
In order to solve this problem, the industry has proposed some new improvements, which are summarized into the following two categories:
(1) the multi-machine redundancy scheme is characterized in that a working group control unit and one or more standby group control units are arranged in an elevator group control system, and once the working group control unit fails in the running process, the working group control unit is switched to the standby group control unit;
(2) the solution of the semi-centralized control mode is that the group control unit is cancelled, each elevator has the group control function, one elevator is taken as the group control unit at a certain moment, and once the current group control unit is found to be in failure, one of the other elevators plays the role of the group control unit according to the priority agreed in advance, for example, the invention patents such as CN201210381192.6, CN201610117907.5 and CN201610119384.8 adopt the solution.
In the above-mentioned improvement, the first solution obviously increases the cost of the equipment additionally, and in order to ensure the high reliability of the elevator system, a multi-level backup needs to be selected, which further increases the cost, and details such as hot switching of the group control unit when a fault occurs also have certain technical difficulties. Although the second solution avoids the cost problem of extra backup, as the group control unit of the elevator group control system, it needs to undertake the dispatching responsibility, which results in the increase of software and hardware requirements and the extra cost, and at the same time, the connection mode between the elevators becomes complicated, and the fault monitoring and the timely switching when the fault occurs are still complicated.
In the two improved schemes, the elevator group control system still belongs to a centralized control mode essentially when in operation, and equipment serving as a group control unit completes communication with all elevators and obtains states of the elevators to analyze, so that a scheduling decision is made, the workload is high, the requirement on hardware is still high, and the cost is also high.
Disclosure of Invention
The invention aims to solve the technical problem of providing an elevator group control system and method based on a multi-agent competition mode, and solving the problems of low reliability and high requirements on software and hardware of the conventional elevator group control system.
In order to solve the technical problems, each elevator control system in the elevator group control system is defined as an autonomous intelligent agent, the intelligent agent obtaining a calling signal generates bidding information according to a preset protocol and sends the bidding information to other intelligent agents in a cooperative range, each intelligent agent makes a response and bids for the bidding information, and finally the intelligent agent with the optimal bidding price obtains a service authority to provide services for passengers;
each agent comprises a user interface unit, a bidding unit, a communication unit and a control unit, wherein:
a user interface unit for interacting with a passenger;
the communication unit is used for carrying out information transmission with other intelligent agents in the cooperative range;
the bidding unit is used for generating all information in the bidding process and bidding with other intelligent agents in the cooperative range;
and the control unit is a service execution unit and operates the plan and executes the service according to the call form after the elevator wins the bid.
Preferably, the agent further comprises a self-learning unit which self-learns according to all data in the bidding process and the service process and improves the bidding mode of the bidding unit.
In the system, the user interface unit and the communication unit carry out information bidirectional transmission, the bidding unit, the user interface unit and the self-learning unit carry out information bidirectional transmission, the bidding unit carries out information transmission to the control unit, and the bidding unit and the control unit carry out information transmission to the self-learning unit together.
The invention also provides an elevator group control method based on a multi-agent competition mode, each elevator control system in the elevator group control system is defined as an autonomous agent, the agent obtaining the call signal generates bidding information and sends the bidding information to other agents in a cooperative range according to a preset protocol, each agent makes a response and bids aiming at the bidding information, and finally the agent with the optimal bidding price obtains service authority to provide service for passengers.
In the method, the triggering condition for the intelligent agent to enter the bidding process is that a passenger sends a calling signal or a reservation signal through a calling device or a reservation device connected with the intelligent agent, or the intelligent agent detects that new bidding information exists on the network.
Furthermore, the calling signal or the reservation signal of the passenger is transmitted to the user interface unit of the intelligent body in real time through the communication unit of the intelligent body; after receiving the calling signal, the user interface unit makes standard service request information according to a specified protocol and transmits the service request information to the bidding unit of the intelligent agent; after receiving the service request information, the bidding unit calculates the bidding price and generates bidding information aiming at the service request information, and then transfers the bidding information to the communication unit through the user interface unit; and after receiving the bidding information, the communication unit sends the bidding information to the network according to a protocol agreed in advance.
Furthermore, the communication unit of the intelligent agent constantly monitors the network where the intelligent agent is located, and after receiving the bid inviting information, the communication unit transmits the bid inviting information to the bidding unit of the intelligent agent through the user interface unit of the intelligent agent; the bidding unit calculates the bidding price aiming at the bidding information and generates the bidding information, and then transfers the bidding information to the communication unit through the user interface unit; and after receiving the bidding information, the communication unit sends the bidding information to the network according to a protocol agreed in advance. The method comprises the following specific steps:
step one, after a communication unit of a certain intelligent agent receives bid inviting and bidding information and transmits the bid inviting and bidding information to a bid inviting unit, the bid inviting unit judges whether the communication unit has service capability to the bid inviting information, namely whether the communication unit meets the bid inviting requirement, if the communication unit meets the bid inviting requirement, the second step is carried out, otherwise, the bid price is set as the appointed worst price, and the seventh step is carried out;
secondly, the bidding unit generates a bidding price according to the self condition and a set bidding price algorithm;
thirdly, the bidding unit compares the generated bidding price with the current optimal price in the bidding information, if the generated bidding price is superior to the current optimal price, the fourth step is carried out, and if the generated bidding price is not superior to the current optimal price, the fifth step is carried out;
fourthly, replacing the current optimal price and the optimal price bidder by the bidding unit, and entering a seventh step;
fifthly, the bidding unit compares the generated bidding price with the current candidate price in the bidding information, and if the generated bidding price is superior to the current candidate price, the operation enters a sixth step;
sixthly, the bidding unit replaces the current candidate price and the candidate price bidder, and the seventh step is entered;
and seventhly, the bidding unit updates bidding information including the bidder list, transmits the bidding information to the communication unit through the user interface unit, and the communication unit transmits the bidding information of the intelligent agent to the network.
In the above method, the termination condition of the bidding process is that all agents within the collaborative range have bid or that the time interval between the current time and the bidding start time has exceeded the set maximum bidding period. Furthermore, when finding that the bidding meets the termination condition, each intelligent agent sets the bidding state in the bidding information to be completed, and broadcasts the bidding state to all intelligent agents in the cooperative range. The method comprises the following specific steps:
step one, a bidding unit of the intelligent agent compares a bidder list in the bidding information with a member list, if all the intelligent agents bid, the fourth step is carried out, otherwise, the second step is carried out;
secondly, the bidding unit calculates the bidding elapsed time, namely the time interval between the current time and the bidding starting time;
thirdly, the bidding unit judges whether the bidding elapsed time exceeds the longest bidding period, if so, the fourth step is carried out, otherwise, bidding information is made and transmitted to the network where the bidding information is located;
and fourthly, the bidding unit sets a bidding ending mark for the bidding state in the bidding information, transmits the bidding ending information to the communication unit through the user interface unit, and sends the bidding ending information to all the intelligent agents in the cooperative range through the communication unit.
In the method, after the bidding process is finished, the intelligent agent confirms whether the intelligent agent is a successful bidder, if the intelligent agent is the successful bidder, the bidding unit of the intelligent agent informs the control unit of the intelligent agent of successful bidding information, and simultaneously generates supporting information and sends the supporting information to each online intelligent agent in a cooperative range; if the intelligent agent is not a successful bidder, further determining whether the intelligent agent is a candidate successful bidder, if so, starting timing to wait for the successful bidder to issue the carrying information, if the bidding unit of the intelligent agent cannot receive the carrying information of the successful bidder within the specified time, switching the intelligent agent to the successful bidder and informing the control unit of the successful bidding information, simultaneously issuing the carrying information, and explicitly recording the carrying service of the candidate successful bidder.
Furthermore, after receiving the bid-winning information, the control unit of the intelligent agent forms an execution plan and executes service, meanwhile, all the intelligent agents mark the bid-winning information into a service registration table held by the bidding unit, and after the intelligent agent executing the service completes one-time service, the control unit of the intelligent agent informs the bidding unit, the bidding unit generates service completion information, and the service completion information is sent to all the intelligent agents through the user interface unit and the communication unit; the service registration table is used for recording all bidding information and final acceptors in the elevator group control system.
The bidding unit of the intelligent agent traverses the service registration table according to a specific period, generates re-bidding information when bidding information capable of being re-bid is found in the service registration table, and requires the communication unit to send the re-bidding information to the current receiver of the service through the user interface unit; after the intelligent agent as the current receiver receives the re-bidding information, the bidding unit re-evaluates the service, if the bidding price is still better than that of the re-bidder, the result is the maintenance of the current situation, otherwise, the re-bidder bids the bid, and the bidding unit of the current receiver generates the re-bidding information according to the result and transmits the re-bidding information to the re-bidder.
When one intelligent agent wins a bid but can not fulfill the carried service, the bidding unit of the intelligent agent generates fault information and issues the fault information to the network, after each intelligent agent receives the fault information, the bidding unit checks the service registration table, revises the record of all the fault intelligent agents in the table as winning bidders, replaces the winning bidders and the winning bid prices with candidate price and candidate price bidders, informs the control unit of the newly elected winning bidder of the service, and the control unit lists a service plan and executes the service.
The invention has the beneficial effects that:
1. the existing centralized control mode or the improved centralized control mode is thoroughly changed, and a completely distributed mode is adopted, so that the elevator group control system does not depend on a single group control unit any more, the group control failure caused by the failure of the group control unit is avoided, the group control mode is changed from the scheduling work under limited resources to the cooperative work of independent individual competition tasks, and the expansibility of the system is greatly improved;
in the prior art, the number of control consoles of a group control system is mainly limited by the number of hardware interfaces serving as group control units and the computing capacity, but in the technical scheme of the invention, as long as network access is allowed, the group members can be arbitrarily increased, and the computing load of each member is not increased;
2. the bidding mode is adopted to compete for the service authority form to work cooperatively, meanwhile, in order to solve some special specification calls or specific service requirements, a bidding qualification system is set, only an intelligent agent meeting the bidding requirement can bid, and the realization of the special specification is facilitated;
3. the bidding result does not depend on a single group control unit to carry out service distribution, but each intelligent agent determines according to the protocol, and the successful bidding acceptor which obtains the service authority publishes the result according to the protocol;
4. each agent has the rule of re-determining the receiver aiming at the possible abnormal condition, and simultaneously introduces a mechanism of carrying out secondary competition at any time, and can carry out replacement aiming at some unsuitable receivers according to the condition change, so that the group control system has an error correction mechanism;
5. because a multi-agent competition mode is adopted, the calculation load is dispersed, and each agent independently self-learns according to the scheduling process and the operation process, so that the realization is easy and the updating is carried out, and the estimated result and the actual result of time-casting are closer to each other by gradual optimization;
6. because the centralized control mode is cancelled, extra group control unit equipment is not needed, the equipment cost is saved, and the fault points are reduced; meanwhile, the service is dispersed in each elevator, the running frequency is dispersed, the group control equipment cannot be frequently replaced, and the elevator group control system is more beneficial to economy.
Drawings
FIG. 1 is a block diagram of the structure of an agent of the present invention;
fig. 2 is a schematic diagram of the networking of an elevator group of the present invention;
FIG. 3 is a flowchart illustrating operation of a bidding unit within an agent of the present invention;
FIG. 4 is a flow chart of bid information generation according to the present invention;
FIG. 5 is a flowchart illustrating the decision of bid completion according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
On the basis of the existing elevator control system which operates independently, the functions of the elevator control system are expanded, so that the elevator control system has a functional module which can make corresponding decisions for the received service request, and each elevator control system can exchange information with other elevator control systems coordinated with the elevator control system.
The service request is an action that a passenger calls an elevator in a landing through an operation device (which can be a landing button, a destination floor selection device, a mobile phone and any other device capable of sending a call request) and requires the elevator to come for service, and in general, the service request generates a corresponding call signal (or called a car call signal). In the known elevator group control system, the call signal is transmitted to the group control unit via the communication unit of the elevator.
In the elevator group control system based on the multi-agent competition mode, each elevator control system is an independent agent, and the agents obtain the service authority through the competition mode and then serve passengers. In general, a call signal from a passenger at a landing is transmitted to an elevator connected to an operation device (i.e., a calling device). In the invention, the car calling device on each landing is connected with one elevator control system, the elevator control system which obtains the call signal packages the call signal according to a preset protocol, makes bid inviting information and sends the bid inviting information to other elevator control systems in a cooperative range, each elevator control system responds after receiving the bid inviting information, bids aiming at the call signal, and finally obtains service authority by an intelligent body with the optimal bid, thereby providing service for passengers.
Each agent includes a user interface unit (i.e., UI unit), a bidding unit, a communication unit, and a control unit, as shown in fig. 1, wherein:
a user interface unit for interacting with passengers, mainly including receiving call signals of passengers, performing service prompt (informing passengers which elevator is served by the passenger through a display screen, providing elevator state of service, such as position, direction, etc.), etc., and directly connecting a landing calling device or an in-car calling device and a service prompt device (such as a display screen, etc.) through a wired manner or a wireless manner;
the communication unit is used for carrying out information transmission with other intelligent agents in the cooperative range;
the bidding unit is used for generating all information in the bidding process, including bidding information, bidding price and the like, and simultaneously transmitting the information to the communication unit through the user interface unit and bidding with other intelligent agents in the cooperative range;
the control unit, which is an elevator controller in the conventional sense, is a service execution unit, and when a bid is successful for a certain elevator control system, which is an agent, the control unit in the agent forms an execution plan according to the call form and executes the service to provide the passenger with the desired service.
The user interface unit is in a logic meaning and comprises an operation receiving part and an information prompting part, wherein the operation receiving part comprises a traditional landing button, a car button, a novel mobile phone application program (APP), a card swiping machine, even a biological identification terminal and the like which can essentially output a call calling signal, and the information prompting part is a device for feeding elevator service information back to a user and comprises a traditional landing direction and floor display, a display screen in a car, a mobile phone screen and the like. The user interface unit receives the operation of the passenger, generates a call signal to the subsequent unit, displays some current selected services or other humanized information in the process of bidding of the subsequent unit, and displays the bidding result to the passenger after the bidding result is generated. For example, when the passenger sends a call signal and no bidding result is generated, the user interface unit may display information of "you choose to go to N floors, dispatch a suitable elevator and please wait", and after the bidding result is generated, the user interface unit may display elevator service information of "m number of elevators served for you are arriving and please wait to m number of sides".
As shown in fig. 1, the user interface unit and the communication unit perform information bidirectional transmission, the bidding unit and the user interface unit perform information bidirectional transmission, and the bidding unit performs information transmission to the control unit.
In the invention, a plurality of elevators which work cooperatively are called as an elevator group, correspondingly, a plurality of elevator control systems form an elevator group control system, and the communication mode of the elevator group control system can take various forms as long as the requirement that all the elevator control systems finish the bidding process within the appointed time is met. For example, a wired connection (various structures such as a bus type, a mesh type, and a star type) may be physically used, or a wireless connection may be used; the communication form can be event-driven or polling mode, and the skilled person can carry out the equipment according to the actual need. Typically, the passenger operates the call button at the landing and the system needs to feed the results back to the passenger within seconds, as shown in fig. 2 for a bus-type networking architecture.
The elevator group control system responds to the call signals of passengers and feeds back results, mainly comprises a bidding process and a successful bid execution process, and also relates to abnormal condition processing such as re-bidding, service receiver failure and the like in the process. The following is a detailed description of each process.
In the present invention, the bidding unit of each agent can be triggered to enter into a working state (i.e. bidding process) by two factors, one is that the passenger sends a call (or reservation) signal through the car calling (or reservation) device connected with the agent, and the other is that the agent detects that there is new bidding information on the network through the communication unit, and the related flow is as shown in fig. 3.
The specific process that the intelligent agent enters the working state through the first factor is as follows: the passenger calls (or reserves) the elevator through a car calling device or a reserving device in the landing, for example, the passenger calling operation generates a call signal by using an operation button, a floor selector, a mobile phone APP and the like of the landing, and the call signal is transmitted to a user interface unit of an intelligent agent corresponding to the car calling device in real time through a communication unit. After receiving the call signal, the user interface unit makes standard service request information (namely, a package call signal) according to a specified protocol, and transmits the service request information to the bidding unit of the intelligent agent. After receiving the service request information, the bidding unit calculates the bidding price for the service request information and generates bidding information, and then transfers the bidding information to the communication unit via the user interface unit. And after receiving the bidding information, the communication unit sends the bidding information to a network (the network where all the cooperative intelligent agents are located) according to a protocol agreed in advance.
The flow of the intelligent agent entering the working state through the second factor is similar to the flow of entering the working state through the call triggering, the difference is that the communication unit of the intelligent agent constantly monitors the network where the intelligent agent is located, the communication unit transmits bidding information to the bidding unit after receiving the bidding information, the subsequent flow is the same as the call triggering, namely the bidding unit calculates the bidding price, generates the bidding information, transmits the bidding information to the communication unit through the user interface unit and releases the bidding information by the communication unit. The communication unit monitoring network has two main functions: (1) if the network has data, receiving and transmitting the data to a subsequent corresponding unit for processing; (2) and obtaining the network sending authority to send data. If the networking uses a general protocol such as TCP, the collision avoidance is a default function, and reference may also be made to the implementation of the collision avoidance function, which is already the existing mature technology of network communication and will not be described in detail here.
The bidding information includes a bidding number, an issuer, a call attribute (departure floor, front/back, direction, destination floor, special specification), a current optimal price, i.e., a bidder number, a current candidate price, and a bidder number, as shown in table 1, but is not limited thereto, and those skilled in the art may add or subtract related information as needed.
TABLE 1 Bidding information Structure
The bidding of an agent is the whole process from receiving bidding information to generating the own bidding price, namely the step of generating the bidding information in fig. 3. No matter which factor triggers the intelligent agent to enter the working state, the bidding starts from the discovery of bidding information, but the difference is that the specific process that the bidding unit of the intelligent agent triggers the working state by the first factor is that the bidding unit of the intelligent agent generates the bidding information after receiving the service request information and automatically enters the bidding link, and the specific process that the bidding unit of the intelligent agent triggers the working state by the second factor is that the bidding information is discovered on the network and then the bidding link is entered. The specific steps are shown in fig. 4 and are explained as follows:
step one, after a communication unit of a certain intelligent agent receives bid inviting and bidding information and transmits the bid inviting and bidding information to a bid inviting unit, the bid inviting unit judges whether the communication unit has service capability to the bid inviting information, namely whether the communication unit meets the bid inviting requirement, if the communication unit meets the bid inviting requirement, the second step is carried out, otherwise, the bid price is set as the appointed worst price, and the seventh step is carried out;
secondly, the bidding unit generates a bidding price according to the self condition and a set bidding price algorithm;
thirdly, the bidding unit compares the generated bidding price with the current optimal price in the bidding information, if the generated bidding price is superior to the current optimal price, the fourth step is carried out, and if the generated bidding price is not superior to the current optimal price, the fifth step is carried out;
fourthly, replacing the current optimal price and the optimal price bidder by the bidding unit, and entering a seventh step;
fifthly, the bidding unit compares the generated bidding price with the current candidate price in the bidding information, and if the generated bidding price is superior to the current candidate price, the operation enters a sixth step;
sixthly, the bidding unit replaces the current candidate price and the candidate price bidder, and the seventh step is entered;
and seventhly, the bidding unit updates bidding information including the bidder list, transmits the bidding information to the communication unit through the user interface unit, and the communication unit transmits the bidding information of the intelligent agent to the network. In order to avoid conflict of bids of a plurality of agents in a network, a standard network protocol such as TCP can be adopted, the network conflict is automatically solved by an underlying protocol, if no network protocol exists, a random number is generated to wait for a time interval and then access is carried out again, the network conflict can be avoided because the random numbers obtained by different agents are different and the waiting intervals are different, and the technology belongs to a mature means of network communication and is not detailed herein.
During the bidding process, the bidding unit of each agent determines whether the bidding process should be completed. In the present invention, the bidding process may be terminated by two conditions, the first being that all agents within the collaborative scope have bid, and the second being that the time interval between the current time and the bidding start time has exceeded the set maximum bidding deadline.
When finding that the bid inviting meets the termination condition, each agent ends the bid inviting state in the bid inviting and bidding information and broadcasts the bid inviting state to all agents in the cooperative range, and the process is shown in fig. 5 and specifically includes the following steps:
step one, the bidding unit compares the bidder list with the member list stored in the bidding unit, if all the agents have bid, the fourth step is carried out, otherwise, the second step is carried out;
secondly, the bidding unit calculates the bidding elapsed time, namely the time interval between the current time and the bidding starting time;
thirdly, the bidding unit judges whether the bidding elapsed time exceeds the longest bidding period, if so, the fourth step is carried out, otherwise, bidding information is made and transmitted to the network where the bidding information is located;
and fourthly, the bidding unit sets a bidding ending mark for the bidding state in the bidding information, transmits the bidding ending information to the communication unit through the user interface unit, and sends the bidding ending information to all the intelligent agents in the cooperative range through the communication unit.
When the bidding unit of a certain intelligent agent finds that the bidding termination condition is met, generating bidding termination information (namely bidding status and bidding termination of the bidding information) and sending the generated bidding termination information to each online intelligent agent in the cooperation range, wherein the bidder at the current optimal price in the bidding information is the successful bidder, and the successful bidder provides service for the passenger. After a communication unit of a certain intelligent agent receives bidding ending information, the bidding ending information is transmitted to a bidding unit through a user interface unit, the bidding unit confirms whether the intelligent agent is a successful bidder (namely, the bidding unit confirms whether the number of an optimal price bidder is the number of the intelligent agent), if the intelligent agent is the successful bidder, the bidding unit informs a control unit of successful bidding information, and simultaneously generates carrying information and transmits the carrying information to each online intelligent agent in a cooperative range; if the intelligent agent is not a successful bidder, whether the intelligent agent is a candidate successful bidder is further confirmed, if so, timing is started, and the intelligent agent waits for the successful bidder to issue the receiving information, if the bidding unit of the intelligent agent cannot receive the receiving information of the successful bidder within a specified time, the intelligent agent switches the intelligent agent to the successful bidder and notifies the control unit of the successful bidding information, and simultaneously issues the receiving information, and clearly describes the receiving service (the candidate bidder described in the bidding information) of the candidate successful bidder.
And after receiving the bid-winning information, the control unit forms an execution plan and executes the service. All agents mark the current winning information into a held service registration table, which is held and managed by a bidding unit and is used for recording all information of bidding information, final acceptors and the like in the elevator group control system.
When one intelligent body finishes one service, the control unit informs the bidding unit, the bidding unit generates service finishing information and sends the service finishing information to all the intelligent bodies through the user interface unit and the communication unit, all the intelligent bodies delete the service from the service registration list after receiving the information and move the service to a history record (the history record is an independent area in the bidding unit and stores each piece of bidding information) for storage, and the service finishing information mainly comprises service process related data, such as time from receiving the bidding information to service starting, service providing time and the like.
During the bidding process, some abnormal situations, such as re-bidding, failure of service carrier, etc., occur inevitably.
During operation, the bidding unit of the intelligent agent traverses the service registration table according to a specific period to find the service requirement expected to be bid again. When bidding information that can be newly bid is found in the service registry, the bidding unit generates re-bid information and requests the communication unit to transmit the re-bid information to the current recipient of the service via the user interface unit. After receiving the re-bidding information, the intelligent communication unit as the current receiver transmits the re-bidding information to the bidding unit via the user interface unit, the bidding unit re-evaluates the service, if the bidding price is still superior to that of the re-bidder, the result is the current state maintenance, otherwise, the re-bidder bids the bid, and the bidding unit generates the re-bidding information according to the result and transmits the re-bidding information to the re-bidder. After receiving the re-bidding information, the re-bidder publishes the re-bidding information to all agents if the re-bidder wins (in the same way of publishing the bidding result), and all agents update the service registry.
When one intelligent agent won't fulfill the accepted service, the bidding unit generates fault information and sends it to the network via the user interface unit and communication unit, the fault information includes fault number and fault attribute, the communication unit of each intelligent agent receives the fault information and informs the bidding unit via the user interface unit, the bidding unit checks the service register list, revises all fault intelligent agents in the list as the record of winning bid, replaces the winning bid and winning bid with candidate price and candidate price bidder, the candidate field information remains unchanged, that is, the revised winning bid and candidate of these readjusted services are the same. The newly elected winning bidder (i.e., the original candidate) notifies the control unit of the service, and the control unit lists the service plan and executes the service. In addition, when the intelligent agent finds that services carried by the intelligent agent do not fit to be provided by the intelligent agent, the intelligent agent also issues information of re-bidding, and the process is the same as the process of normal bidding.
In the bidding process, the bid price of the intelligent agent for a certain service request is calculated according to the current state of the intelligent agent and the characteristics of the service according to a certain bid price algorithm, and the calculation result reflects the expectation of the quality of the service provided by the intelligent agent. In general, there are two cases of bid prices, the first is a waiting time class index, where the lower the value the better, such as the boarding time, the average waiting time of passengers during elevator operation, etc., and can also be a simple function of these service or waiting time indices; the second is the reward obtained by the service provided by the agent, the higher the value, the better, in this case, the party who reports high reward wins in the bidding process, usually some indirect index is adopted, the agent is encouraged to take multiple orders, and if the order completion time is lower than the average system period (average of all the agent ties or the history of the agent), the product of the difference part and the total number of the accepted service is used as the reward to the agent according to a certain proportionality coefficient.
The algorithm for calculating the bid price can be an explicit function calculation process, and can also be a calculation model such as a neural network (including a deep neural network). As a specific implementation, reference is made to the solution published in the invention patent application with application number 201310080795.7.
On the basis of the elevator group control system, the intelligent behavior of the intelligent agent is better than the bidding behavior, and the intelligent agent has the advantages that the bidding process and the service process can be learned, and the calculation capability of the bidding price of the intelligent agent is continuously improved. Therefore, the intelligent agent also comprises a self-learning unit which can carry out self-learning according to all data in the bidding process and the service execution process and improve the bidding mode of the bidding unit, and as shown in fig. 1, the bidding unit and the control unit carry out information transmission to the self-learning unit together.
In the invention, the self-learning unit of each intelligent agent independently learns according to the bidding data and the service result in an enhanced learning mode. In the self-learning process, the agent has two states, namely an idle state and a service state. The idle state refers to the state that the current operation plan of the agent does not have any service request which is not completed and is in a static state, and belongs to a task waiting state. The service state refers to that the agent has an incomplete service request in the operation plan or is executing a service process although the incomplete service request does not exist. The process of entering the service state from the idle state and completing all service plans to enter the idle state again is called a scenario (epicode). After each scene is finished, the self-learning unit of the intelligent agent self-learns the data in the scene process.
The goal of the intelligent agent for self-learning is that the predicted data of the bid price and the actual data of the final service completion are consistent as much as possible, namely, the predicted service quality during bid is more and more accurate through self-learning, so that the best bid price is the best service during bid.
The self-learning result of the intelligent agent in one scene is temporarily stored in the self-learning unit, and the bidding price calculation rule is not updated immediately. And after the scenes experienced by the agents reach a certain number (for example, 50), uniformly updating the bid price calculation rules of all the agents, namely synchronously updating the bid price calculation rules of all the agents by using the result learned by one agent. After the self-learning unit of an agent performs self-learning in a specified number of scenes, the self-learning unit sends a learning notice to other agents in a cooperative range in a broadcasting mode (the learning notice is sent by the self-learning unit, sequentially passes through the bidding unit and the user interface unit, and is finally broadcast to a network where the communication unit is located), a learning result is prepared to be issued, the agents feed back that the issuers are ready to receive data after receiving the learning notice, the issuers update the bidding price calculation rules of the agents in the group after issuing the learning result, the bidding price calculation rules are fed back to the issuers after the updating is successful, and the issuers record a successful updater list and version. If a certain agent fails to update successfully, it does not matter, because the updating of the learning result directly covers the old parameters, and the updating can skip the intermediate stage to the latest result.
Each learning update has a unique version number, e.g., the version number is 0 at the beginning of the system, and 1 is added to each update. When an agent prepares to issue a learning result, a new version number is automatically generated according to the existing record when a learning notice is issued, and is published to all members in the group, and each member updates the internal record and maintains the latest version number. The version number cannot be revoked once published regardless of whether the learning result update is successful or not. If the published update version is lower than the current latest version, the update is ignored.
Although in the above description, all agents within the cooperation range collectively update the bid price calculation rule, the present invention is not limited thereto, and the agents may be updated individually. In addition, although all the data information generated by the bidding unit is transmitted to the communication unit through the user interface unit and then is issued by the communication unit, a person skilled in the art can easily improve the data information according to the above principle, so that the information generated by the bidding unit is directly issued through the communication unit without passing through the user interface unit, which is described herein.
The invention thoroughly changes the existing centralized control mode or the improved centralized control mode and adopts the completely distributed mode, so that the elevator group control system does not depend on a single group control unit any more, the group control failure caused by the failure of the group control unit is avoided, the group control mode is basically changed from the scheduling work under limited resources to the cooperative work of independent individual competition tasks, and the expansibility of the system is greatly improved. In the prior art, the number of control consoles of the group control system is mainly limited by the number of hardware interfaces serving as group control units and the computing capacity, but in the technical scheme of the invention, the group members can be arbitrarily increased as long as the network access is allowed, and the computing load of each member is not increased at all.
The invention adopts a bidding mode to compete for the service authority form to work cooperatively, and simultaneously, in order to solve some special specification calls or specific service requirements, a bidding qualification system is set, only an intelligent agent meeting the bidding requirement can bid, thereby facilitating the realization of the special specification; the bidding result does not depend on a single group control unit to carry out service distribution, but each agent decides according to the protocol, and the successful receiver who obtains the service authority publishes the result according to the agreement.
Meanwhile, each agent of the invention has the rule of re-determining the receiver aiming at the possible abnormal condition, and simultaneously introduces a mechanism of carrying out secondary competition at any time, and can replace some unsuitable receivers according to the condition change, so that the group control system has an error correction mechanism; due to the adoption of a multi-agent competition mode, the independent self-learning of each agent is easy to realize, and the agent can independently learn and update according to the running process, so that the expected result and the actual result of time-casting are closer to each other by gradual optimization.
The invention cancels the centralized control mode, so additional group control unit equipment is not needed, the equipment cost is saved, and the fault points are reduced; meanwhile, the service is dispersed in each elevator, the running frequency is dispersed, the group control equipment cannot be frequently replaced, and the elevator group control system is more beneficial to economy.
The present invention has been described in detail with reference to the specific embodiments, which are only the preferred embodiments of the present invention and are not intended to limit the present invention. Equivalent alterations and modifications made by those skilled in the art without departing from the principle of the invention should be considered to be within the technical scope of the invention.
Claims (14)
1. An elevator group control system with a multi-agent competition mode is characterized in that each elevator control system in the elevator group control system is defined as an autonomous agent, the agent obtaining call signals generates bidding information according to a preset protocol and sends the bidding information to other agents in a coordination range, each agent makes a response and bids according to the bidding information, and finally the agent with the optimal bidding price obtains service authority to provide service for passengers;
each agent comprises a user interface unit, a bidding unit, a communication unit and a control unit, wherein:
a user interface unit for interacting with a passenger;
the communication unit is used for carrying out information transmission with other intelligent agents in the cooperative range;
the bidding unit is used for generating all information in the bidding process and bidding with other intelligent agents in the cooperative range;
and the control unit is a service execution unit and operates the plan and executes the service according to the call form after the elevator wins the bid.
2. The multi-agent competitive elevator group control system of claim 1 wherein the agents further comprise a self-learning unit that self-learns and improves bidding patterns of the bidding units based on all data in the bidding process and the servicing process.
3. The multi-agent competition mode elevator group control system according to claim 1, wherein the user interface unit performs bidirectional information transmission with the communication unit, the bidding unit performs bidirectional information transmission with the user interface unit, and the bidding unit performs information transmission to the control unit.
4. The multi-agent competitive elevator group control system as recited in claim 2, wherein the user interface unit communicates with the communication unit bidirectionally, the bidding unit communicates with both the user interface unit and the self-learning unit bidirectionally, the bidding unit communicates with the control unit, and the control unit communicates with the self-learning unit.
5. An elevator group control method based on a multi-agent competition mode is characterized in that each elevator control system in an elevator group control system is defined as an autonomous agent, the agent obtaining a call signal generates invitation information according to a preset protocol and sends the invitation information to other agents in a cooperative range, and each agent responds to the invitation information and bids;
the method comprises the steps that the end condition of a bidding process is that all intelligent agents in a cooperative range bid or the time interval between the current time and the bid inviting start time exceeds the set longest bid inviting period, and when each intelligent agent finds that the current bid inviting meets the end condition, the bid inviting state in bid inviting information is placed and bid inviting is finished and is broadcasted to all intelligent agents in the cooperative range;
after the bidding process is finished, the intelligent agent confirms whether the intelligent agent is a successful bidder, and if the intelligent agent is the successful bidder, the intelligent agent obtains service authority to provide service for passengers; if the agent is not the successful bidder, further confirming whether the agent is a candidate successful bidder, if so, starting timing, waiting for the successful bidder to issue the carrying information, and if the agent fails to receive the carrying information of the successful bidder within the designated time, switching the agent to the successful bidder and providing service for the passenger.
6. The group control method for elevators based on multi-agent competition mode of claim 5, wherein the triggering condition for the agent to enter the bidding process is that the passenger sends a call signal or a reservation signal through a car calling device or a reservation device connected with the agent, or the agent detects that there is new bidding information on the network.
7. The multi-agent competition mode-based elevator group control method according to claim 6,
the calling signal or the reservation signal of the passenger is transmitted to the user interface unit of the intelligent body in real time through the communication unit of the intelligent body;
after receiving the calling signal, the user interface unit makes standard service request information according to a specified protocol and transmits the service request information to the bidding unit of the intelligent agent;
after receiving the service request information, the bidding unit calculates the bidding price and generates bidding information aiming at the service request information, and then transfers the bidding information to the communication unit through the user interface unit;
and after receiving the bidding information, the communication unit sends the bidding information to the network according to a protocol agreed in advance.
8. The multi-agent competition mode-based elevator group control method according to claim 6,
the communication unit of the intelligent agent constantly monitors the network where the intelligent agent is located, and after receiving the bid inviting information, the communication unit transmits the bid inviting information to the bidding unit of the intelligent agent through the user interface unit of the intelligent agent;
the bidding unit calculates the bidding price aiming at the bidding information and generates the bidding information, and then transfers the bidding information to the communication unit through the user interface unit;
and after receiving the bidding information, the communication unit sends the bidding information to the network according to a protocol agreed in advance.
9. The multi-agent competition mode-based elevator group control method according to claim 7 or 8, characterized by comprising the following steps:
step one, after a communication unit of an intelligent agent receives bid inviting and bidding information and transmits the bid inviting and bidding information to a bid bidding unit, the bid bidding unit judges whether the communication unit has service capability on the bid inviting information, namely whether the communication unit meets the bid inviting requirement or not, if the communication unit meets the bid inviting requirement, the second step is carried out, otherwise, the bid price is set as the worst agreed price, and the seventh step is carried out;
secondly, the bidding unit generates a bidding price according to the self condition and a set bidding price algorithm;
thirdly, the bidding unit compares the generated bidding price with the current optimal price in the bidding information, if the generated bidding price is superior to the current optimal price, the fourth step is carried out, and if the generated bidding price is not superior to the current optimal price, the fifth step is carried out;
fourthly, replacing the current optimal price and the optimal price bidder by the bidding unit, and entering a seventh step;
fifthly, the bidding unit compares the generated bidding price with the current candidate price in the bidding information, and if the generated bidding price is superior to the current candidate price, the operation enters a sixth step;
sixthly, replacing the current candidate price and candidate price bidder by the bidding unit, and entering a seventh step;
and seventhly, the bidding unit updates bidding information including the bidder list, transmits the bidding information to the communication unit through the user interface unit, and the communication unit transmits the bidding information of the intelligent agent to the network.
10. The elevator group control method based on multi-agent competition mode as claimed in claim 5, wherein the specific steps are as follows:
step one, a bidding unit of the intelligent agent compares a bidder list in the bidding information with a member list, if all the intelligent agents bid, the fourth step is carried out, otherwise, the second step is carried out;
secondly, the bidding unit calculates the bidding elapsed time, namely the time interval between the current time and the bidding starting time;
thirdly, the bidding unit judges whether the bidding elapsed time exceeds the longest bidding period, if so, the fourth step is carried out, otherwise, bidding information is made and transmitted to the network where the bidding information is located;
and fourthly, the bidding unit sets a bidding ending mark for the bidding state in the bidding information, transmits the bidding ending information to the communication unit through the user interface unit, and sends the bidding ending information to all the intelligent agents in the cooperative range through the communication unit.
11. The group control method for elevators based on multi-agent competition mode as claimed in claim 5, wherein after the bidding process is finished, when the agent confirms that the agent is the successful bidder, the agent generates the receiving information and sends the receiving information to each online agent in the cooperative range; if the user is not a successful bidder, the user further confirms whether the user is a candidate successful bidder, and if the user is a candidate successful bidder and the user cannot receive the receiving information of the successful bidder within a specified time, the user switches the user to a successful bidder and notifies the control unit of the successful bidding information, and at the same time, the receiving information is issued and the receiving service of the candidate successful bidder is explicitly described.
12. The group control method of elevators based on multi-agent competition mode as claimed in claim 11, wherein the control unit of the agents receives the bid-winning information, forms an execution plan and executes the service, and simultaneously all the agents mark the bid-winning information in the service registration table held by the bidding unit, and after the agent executing the service completes one service, the control unit of the agent notifies the bidding unit, and the bidding unit generates the service completion information, and transmits the service completion information to all the agents via the user interface unit and the communication unit; the service registration table is used for recording all bidding information and final acceptors in the elevator group control system.
13. The group control method of elevators based on multi-agent competition mode as claimed in claim 12, wherein the bidding unit of the agent traverses the service registry according to a specific period, generates re-bid information when bidding information that can be re-bid is found in the service registry, and requests the communication unit to transmit the re-bid information to the current recipient of the service via the user interface unit; after the intelligent agent as the current receiver receives the re-bidding information, the bidding unit re-evaluates the service, if the bidding price is still better than that of the re-bidder, the result is the maintenance of the current situation, otherwise, the re-bidder bids the bid, and the bidding unit of the current receiver generates the re-bidding information according to the result and transmits the re-bidding information to the re-bidder.
14. The group control method of elevators based on multi-agent competition mode as claimed in claim 12, wherein when a bid from an agent fails to fulfill the accepted service, the bidding unit of the agent generates a fault message to be issued to the network, and when each agent receives the fault message, the bidding unit checks the service registration table, revises all the faulty agents in the table as the records of winning bidders, replaces the winning bidders and winning bid prices with candidate price and candidate price bidders, notifies the newly elected winning bidder of the service to the control unit, and the control unit lists the service plan and executes the service.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810736500.XA CN108910630B (en) | 2018-07-06 | 2018-07-06 | Elevator group control system and method based on multi-agent competition mode |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810736500.XA CN108910630B (en) | 2018-07-06 | 2018-07-06 | Elevator group control system and method based on multi-agent competition mode |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108910630A CN108910630A (en) | 2018-11-30 |
| CN108910630B true CN108910630B (en) | 2020-09-29 |
Family
ID=64425033
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810736500.XA Active CN108910630B (en) | 2018-07-06 | 2018-07-06 | Elevator group control system and method based on multi-agent competition mode |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108910630B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021214987A1 (en) * | 2020-04-24 | 2021-10-28 | 三菱電機株式会社 | Elevator monitoring system and elevator system |
| CN112591568B (en) * | 2020-12-15 | 2021-11-02 | 海南大学 | Device sharing method for fusion of data and information facing intention calculation and reasoning |
| CN112794172B (en) * | 2021-01-13 | 2022-04-01 | 海南大学 | Data portrait and information portrait value alternating current conversion oriented carrying method and system |
| CN114326608A (en) * | 2021-11-30 | 2022-04-12 | 云南昆船智能装备有限公司 | AGV group system based on multi-agent |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101525099A (en) * | 2009-03-10 | 2009-09-09 | 中山大学 | Device for mutually coordinating elevator equipment in elevator set |
| CN102897613A (en) * | 2012-10-09 | 2013-01-30 | 苏州默纳克控制技术有限公司 | Elevator intelligent group control system and method |
| CN102976170A (en) * | 2011-09-05 | 2013-03-20 | 东芝电梯株式会社 | Group management control device of elevator |
| JP2017154839A (en) * | 2016-02-29 | 2017-09-07 | フジテック株式会社 | Group management control device and group management system for elevators, and elevator system |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE3763766D1 (en) * | 1986-07-07 | 1990-08-23 | Inventio Ag | SYSTEM FOR REMOTE MANAGEMENT OF ELEVATOR SYSTEMS. |
| JP2624816B2 (en) * | 1989-02-14 | 1997-06-25 | 株式会社東芝 | Elevator control device |
| JP4494696B2 (en) * | 1999-10-21 | 2010-06-30 | 三菱電機株式会社 | Elevator group management device |
| CN103092143A (en) * | 2011-11-08 | 2013-05-08 | 天津长城科安电子科技有限公司 | Centralized control of building intelligent system |
| CN107662861A (en) * | 2016-07-31 | 2018-02-06 | 刘凤娇 | A kind of elevator dispatching system |
| CN106453009B (en) * | 2016-11-16 | 2020-03-27 | 四川省亚丁胡杨人力资源集团有限公司 | Service control system of intelligent community |
-
2018
- 2018-07-06 CN CN201810736500.XA patent/CN108910630B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101525099A (en) * | 2009-03-10 | 2009-09-09 | 中山大学 | Device for mutually coordinating elevator equipment in elevator set |
| CN102976170A (en) * | 2011-09-05 | 2013-03-20 | 东芝电梯株式会社 | Group management control device of elevator |
| CN102897613A (en) * | 2012-10-09 | 2013-01-30 | 苏州默纳克控制技术有限公司 | Elevator intelligent group control system and method |
| JP2017154839A (en) * | 2016-02-29 | 2017-09-07 | フジテック株式会社 | Group management control device and group management system for elevators, and elevator system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108910630A (en) | 2018-11-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108910630B (en) | Elevator group control system and method based on multi-agent competition mode | |
| WO2021004115A1 (en) | Combined dispatching optimization method for manned buses and autonomous buses | |
| CN104340784B (en) | Apparatus for controlling elevator | |
| CN111731958A (en) | Method for robot to take elevator in queue, elevator internet of things device and system | |
| CN101528578A (en) | Elevator cross-dispatching system with inter group relative system response (IRSR) dispatching | |
| JP2012075247A (en) | Control program, charge optimization controller and parking lot system | |
| US20190146484A1 (en) | Vehicle External Communication Method and Apparatus, and Terminal | |
| WO2020170337A1 (en) | Vehicle allocation system and vehicle allocation method | |
| CN110489218A (en) | Vehicle-mounted mist computing system task discharging method based on semi-Markovian decision process | |
| CN110255302A (en) | The control processing method of elevator operation mode | |
| CN102204234B (en) | Call control system, call control apparatus, terminal apparatus, and call control method | |
| CN113268352B (en) | Multi-instruction responsive task collaborative management method for universal service robot | |
| CN111242643B (en) | Real-time updating system for manufacturing service information of miniature annular part | |
| EP3569545B1 (en) | A wireless signal device, system and method for elevator service request | |
| CN112882462A (en) | Cloud traffic management device and method thereof | |
| Nguyen et al. | EdgePV: collaborative edge computing framework for task offloading | |
| CN115439196A (en) | Unmanned vehicle order receiving method and device, storage medium and terminal | |
| CN114444700A (en) | Quantum cloud computing platform job scheduling and resource allocation method | |
| CN111400028B (en) | Load balancing processing method for train management | |
| CN114735553A (en) | Robot autonomous elevator taking method | |
| CN110941404A (en) | Shared printing method and system of laser printer | |
| CN111027880A (en) | Method and system for configuring dispatching personnel of full-automatic operation system | |
| CN112887347A (en) | Dynamic migration method and device for edge calculation in industrial internet | |
| RU2389676C1 (en) | Elevator control device (versions) and elevator control method | |
| CN106483936B (en) | The method and system of automatic monitor task |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |