MXPA00002766A - Order processing apparatus and method - Google Patents

Abstract

An apparatus for processing trading orders, comprising:a central server connectable to a plurality of terminals on which user orders are to be entered. The central server further comprises:communication means for transmitting user orders between said terminals and said central server via a network;first storage means for storing received user orders as an array whose elements define a particular first resource ordered by a particular user;the first storage means is also to be used for storing the resources that may be offered by users for exchange against the first resource ordered;second storage means for storing an array of coefficients each representing the proportion of a particular order that is to be satisfied;processing means for retrieving the orders from the first storage means, calculating an optimized set of values of the coefficients with respect to at least one predetermined, adjustable constraint and at least one predetermined, adjustable criterion, storing the optimized coefficient values in said second storage means;and output means for communicating the processed orders and their respective coefficients.

Description

APPARATUS AND METHOD OF PROCESSING OF ORDERS DESCRIPTION OF THE INVENTION The present invention relates to an apparatus and method for optimizing the allocation of resources based on the received orders. There is a wide variety of systems in which a group of users and where each user submits orders consisting of a desired amount of a given resource or objective that is determined to exchange certain amounts of other resources. Examples include a computer programming system that assigns computed resources to users or for jobs submitted by users; the plants that generate electricity offer to supply the energy for a distribution system at different costs and generated from different incentives; a computer processor allocates resources such as memory and wide band 1/0 for different internal processes or applications of the set of programs; and financial traders offer to buy and sell the resources or amounts of financial instruments, such as shares or currencies, in exchange for other financial instruments. REF .: 33097 A number of different technical solutions have been used for the previous assignment or equalization problem. An example is the binary equalization in which an offer by a user to sell a particular amount of a resource is matched to one offered by a different user to buy the amount of that instrument. A second example, in the case of computer-programmed tasks, is to allocate a time division of the time processor for each user instead of a revolving base. These solutions experience a number of drawbacks, that is, it is generally not optimal, the so-called allocation of resources. For example, the proportion and demand are not the same as they should be, as in a computer, most CPU cycles are inactive, but a peak load of the operation of the computer is limited by one of the available resources as a processing time, memory, wide band I / O. The programming systems do not necessarily take into account the priority of the tasks, as if they were required in real time or could be a processed batch, and assigning a time division to each user in rotation is simply a compromise. Equalization of orders is ineffective where a system of binary equalization the order dimension must be matched so that a larger order can never be matched unless it is interrupted under a number of smaller orders. Equalization is also inefficient because in a binary equalization system the equations involve more than 2 instruments that can not be found in general where optimal matches are prevented from being discovered. In a financial market this can lead to non-liquidity which in turn, and intuitively against it can lead to the problem of market volatility. The present invention seeks to mitigate at least partially some or all of the above problems. According to the present invention an apparatus for the processing of commercial orders is provided, the apparatus comprises a central server connectable to a plurality of terminals where the user's commands will be entered, wherein the central server comprises: means of communication to receive the user's commands from the terminals through a network; a first storage means for storing the received user's commands as a series whose elements define a particular first resource ordered by a particular user; a second storage means for storing a series of each of the coefficients representing the proportion of a particular order that is satisfied; and processing means for retrieving the orders of the first storage means, to calculate an optimized group of values of the coefficients with respect to at least one adjustable restriction, predetermined, and to at least one adjustable, predetermined criterion, and storing the optimized coefficient values in the second storage means, the communication means are also for transmitting the processed commands and their respective coefficients. According to . A further aspect of the present invention provides a method for processing commercial orders comprising the steps of: receiving from users the orders each specifying a first particular resource ordered by a particular user and storing them as a series in a first medium of storage; processing the orders retrieved from the first storage means to calculate a group of coefficients each representing the proportion of a particular order that is satisfied; optimizing the values of the coefficients with respect to at least one predetermined adjustable constraint, and to at least one predetermined adjustable criterion; storing the optimized coefficient values in a second storage medium; and generate the processed orders and their respective coefficients. The embodiments of the present invention can produce optimized solutions for equalizing orders for resources. The solutions are not sensitive to the dimensions of the particular orders received. Equalization can be done in a non-deviant and transparent manner for the particular resources that are commercialized. The embodiments of the present invention that improve the efficiency of a commercial system can alleviate the problem of non-liquidity. The embodiments of the invention will now be described, by means of a single example, with reference to the accompanying drawings in which: Figure 1 shows an apparatus according to the present invention; Figure 2 shows a currently preferred embodiment of an apparatus according to the invention; and Figure 3 is a flow chart illustrating a method according to the invention. Referring to Figure 1, a form of apparatus according to the invention comprises a central server 10 and a plurality of terminals 12 where the user's commands are entered. The terminals 12 can be conventional personal computers (PCs) that properly execute the set of programs, or they can be dedicated commercial terminals. Each terminal 12 is equipped with a communication means 14 as an interface and / or a modulator-demodulator or modem for transmitting the commands from the terminals 12 to the central server 10 via a network 16. The central server 10 is comprised of a number of storage media. These can be devices such as integrated memory circuits or magnetic disks. The different storage means may comprise different regions within a common circuit or disk or perhaps distributed among the various different physical devices. In particular, the central server comprises a first storage means 18 for storing the received user commands as a series whose elements define the quantity of a particular first resource ordered by a particular user and a second storage means 20 for storing a series of coefficients each representing the proportion of a particular order that is satisfied. The central server additionally comprises a processing means 22, such as a central processing unit (CPU) executing the instructions of a particular part of the program set. The processing means 22 is for recovering the commands of the first storage means 18, calculating an optimized group of values of the coefficients with respect to at least one adjustable constraint, predetermined and at least one adjustable, predetermined criterion, and storing the values of the optimized coefficient in the second storage medium 20. A communication means 24, such as an interface and / or modulator-demodulator, is included in the central server 10 for communicating the processed orders and their respective coefficients. The communications between the terminals 12 and the central server 10 via the network 16 are preferably carried out by means of universal standards and protocols, such as TCP / IP and a generally accepted interface is used, for example by browsing on a communications network or the Internet . Communications can be made through a network, such as the Internet, or a communications network in which all users of the exchange system can be linked, and where it is controlled by a central server. The orders entered by the users are automatically passed to the server in real time, and they are matched in the batches (whose length and frequency are controlled in the set of programs by the server) in an optimal sense as explained in this document. A second, currently preferred embodiment of the invention is illustrated in Figure 2. Referring to Figure 2, this embodiment includes all the features shown in Figure 1, numbered with the same reference numbers, and several additional features. The additional features can be used with the apparatus of Figure 1 either individually or in combination. The central server 10 additionally comprises a third storage means 26 for storing a series of data representing the current exchange rate between each traded resource and at least one other resource where the date of the exchange rate can be retrieved by the processing means 22 The processing means 22 can also calculate and update the rate of change in the third storage means 26 based on the flow of orders that are satisfied as described above. The new orders stored in the storage medium 18 also define a second particular resource offered in exchange for the first resource. As shown in Figure 2, at least one of the terminals 12 is connected to the central server 10 by means of the sub-server 28 which adds the commands of the users before transmitting them to the central server 10. In addition a sub-server 28 can be provided and they they can be separated geographically such that sub-server 20 adds the commands of the users in a particular region. The communication means 24 additionally communicates the processed orders and their respective coefficients to an additional apparatus 30 for the liquidation of the orders. The additional apparatus 30 performs the functions such as debit and credit of the user's bank accounts in accordance with the orders that are satisfied.

In the preferred embodiment, the system requires a brokerage house that participates in the order process taking the contrary position of the users with respect to the user's equal orders. In this way the brokerage house will be the formal counterpart for all transactions executed in the system. The matching algorithm ensures that the brokerage house does not take market risks, and that it is subject to the condition, the equalization is optimal with respect to certain criteria explained below. The essential function of optimizing the algorithm that remains at the bottom of the system is to simulate an "infinitely intelligent" brokerage house that is able to search through the order flow available in such a way as to satisfy the limited orders in the system for the optimal extension possible. Since there are several possible criteria to determine, from the point of view of the brokerage house, which is the optimum, the theme assumed in this system is that a number of optimization criteria "fall in cascade". This is done in a way that can be controlled in the set of programs. In other words, the brokerage house (which runs the system) will be able to classify the optimization criteria, from the most to the least important. By providing this classification, the algorithm will first seek the best global equalization within the meaning of the first optimization criterion. You will have to find a group of optimal solutions, then look for a smaller group that is also optimal for the second criterion and so on. The most natural situation recommended by the system is first to optimize the volume (for the maximum portion of the order flow that can be satisfied) and then for the utility of the broker. It is assumed that the broker will earn retributions simply by maximizing the utility of the broker. However, other solutions are possible and can be chosen in real time by the operator or regulator. The system is capable of executing in a hierarchical way, where private users can contract through an intermediary institution, such as a compensation bank, that indicates their credit limits, and determines the coverage of the accounts, that communicates these limits of credit to the central computer, and therefore cause orders to truncate if they exceed the credit limits. The system is also able to deal with orders that arbitrarily involve resource flows as defined in the glossary, for example orders that involve more than two raw instruments, orders that involve contracts to release or receive a specific amount of a raw instrument. at a specific future time and orders that involve options for raw instrument transactions in the future. The restriction of "no risk for the broker" is defined in different ways, depending on the types of instruments involved in the application. Case 1 (all orders involving only raw instruments). The restriction takes the form: the aggregate transactions of the broker in the lot involve non-negative coefficients. In other words, the broker retains only the non-negative amounts in each raw instrument separately. Case 2 (all orders involve only contracts to release or receive specified quantities of raw instruments at specified times). The restriction is that, for the aggregate transactions of the broker in the lot, at no time should the cumulative position of the broker be negative (without counting any position prior to his transaction) of any gross instrument. Case 3 (some or all orders that involve options). The restriction is that under all the scenarios of the year or the partial exercise of the options retained as a result of the equalization by the other participants in the market, the broker may exercise his own options (as a result of the equalization) in such a way that the flows resulting from the raw instruments satisfy the restriction of case 2 above. Now we will analyze the invention in a greater generality. First we describe the case where the orders are general simple derivatives, that is, they are represented by a flow of resources represented by a finite number of resource space elements PF (+) x T as explained in the glossary, which are marketed in the manner described in the glossary. Orders are described by a finite portfolio number of raw resources (such as a currency, stock or in fact anything that has a non-negative value); each portfolio with a value data. These orders describe certain mutually agreed resource flows in the portfolio of the raw instruments (as defined in the glossary). A portfolio is offered to trade against another portfolio. An order represents the willingness to enter into an irrevocable commitment to release the components of the portfolio sold in exchange for the purchased portfolio, on the dates of values specified in the definition of the group of elements that define the simple derivation. The examples of simple derivatives are deposits, guarantees, terms (with 2 dates of values) and changes (with some dates of values arbitrarily) as in the transactions that occur in the financial markets. An order that involves only an exchange of a pair of raw instruments that correspond to a point in PF that has precisely two coordinates without zeroing, one negative and the other positive. The values of these two coordinates represent the amount of a resource that is given up or "sold" in the exchange for the other resource that is purchased or "purchased". All other coordinates in PF are zero, in other words for a simple order only two resources are involved. Normally at present, the situation where the instruments are not raw simply means that the orders have the same value dates. The basic restriction defined by the optimization of the algorithm is the restriction that corresponds to that defined in the discussion of the exchange of raw instruments. These restrictions are as indicated below: First, that the degrees of satisfaction of the orders must remain between 0 and 1. Second, that the return of the broker, at any value date, defined as the sum over all the resource flows in the raw instruments, in all value dates up to and including the value date must be non-negative on all raw instruments. This "non-negative" restriction means that at any time value date, the broker will not have a low position in any appeal instrument. The optimization performed is to optimize a number of successive objective functions of alpha (j), j-1 ... N, where alpha (j) is the degree of acceptance of order j. The first of these functions, in a preferred situation, is the volume defined as the sum of the absolute values of the elements of the alpha matrix. F, where F is the matrix order (whose element (j, k) denotes the quantity of the raw instrument k-th bought or sold in the order j-th, where alpha.F is not the parent product but denotes the matrix resulting from the scalar multiplication of F by line to alpha line (ie alpha (j)) that multiplies the line F (j, 1), ... F (j, k), J-1 ... N), and where the dates of values are ignored).

An order that involves only one exchange of raw instruments, such as the one described in the first part of this specification, is a special case that corresponds to an order in which a simple resource is offered against another simple resource. If the order satisfies certain alpha grades (a number between 0 and 1) where the portion of the transaction has been entered into, and each participant in the transaction is required to receive or release (according to the signal from the order component) ) the raw resource portfolio specifies on its specified value date. Now we return to the case where the orders not only involve raw instruments or simple derivatives, but also options. As already explained, the difference between derivatives and simple options is that, in the latter case, the flow of resources is made in reserve of one of the participants in the option transaction (the retention of a long position in the option) . In this case, the broker must ensure that, without taking into account that other participants will do so, on the last value date that occurs in the order flow, he will not have exposure, that is, he will hold the non-negative positions in all the instruments. He will do this by means of a strategy that is defined inductively, by a number of stages equal to the number of value dates that occur in the low and long positions in the options. The induction will begin with the last option value date. The runner will calculate the limits of the region saved as in the case of simple derivatives and calculate the acceptance coefficients accordingly. Through induction, he will then choose the limits of all previous regions and hence the acceptance coefficients of the option orders with the first values dates, until he finally finds the appropriate region (and hence the coefficients) for the first option date that occurs in the order flow. The method according to one embodiment of the invention will now be described with reference to the flow diagram of Figure 3. In the Sl stage the orders are received from the users in a subserver and aggregated and transmitted to the central server. In step S2 the central server receives the commands directly from the users and adds them through the server in the stage Sl and the incomplete orders from the previous lots. The central server forms the orders received in the batch. The end of a lot is determined either by the volume of orders that exceed a threshold or a fixed time that has elapsed since the previous batch. In step S3 the command batch is stored in a first storage medium . In step S4 the processor of the central server retrieves the data of the exchange rate of a third storage means and also retrieves the command batch of the first storage means. In step S5 the orders are processed to calculate a group of parameters that define the optimization problem. The coefficients are then optimized in step S6. The optimization is subject to the restrictions since the coefficients must be less than or equal to 1 and greater than or equal to 0 and ensure that the broker is not exposed to any risk. The coefficients are optimized with respect to a particular criterion, such as, maximize the total volume to satisfy the orders and maximize the utility of the brokers. The optimization is done by a module of the set of programs. If an optimal solution is found, the system proceeds to step S7. If a current period of time elapses without finding an optimal solution the system proceeds to step S8.

In step S7 the optimization routine generates the data representing the optimized coefficients. In step S8 the optimization routine generates the data representing a sub-optimal group of coefficient values. In step S9 the data represent the coefficient values generated in step S7 or in step S8 are stored in a second storage means. If all the optimization criteria have been applied, the system proceeds to the SIO stage, on the other hand the system returns to the S6 system to optimize the equalization with respect to the following criterion. A number of different optimization criteria are successively applied (usually 2, which corresponds to maximizing the volume and maximizing the utility of the broker). In the SIO stage, the resulting preceded orders and their coefficients are generated. This output is communicated to the users and to a mechanism for the liquidation of the orders. In the Sil stage the central server processor calculates the new exchange rate data for resource transactions based on the satisfied order flow. The new exchange rate data is used to update itself in the third storage medium. The data of the type of change can also be communicated to the users of the system. In step S12 the orders that are not filled, either "completely or partially, and which have not been separated by the user who presented them and for which a specific period of time has not elapsed since they were submitted to return to be processed in the next batch with any of the new orders received from the users in step S2 The sources of algorithms to solve the optimization problems whereby this method reduces the problem of optimal matching are: NK Karmarkar. polynominal-time algorithm for linear programming, Combinatorica, 4: 373-395, C. Ross T. Terlaky, J-Ph Val Theory and Algorithms for Linear Potimization, An Interior Point Approach, J. Wiley, 1997. B. Jansen C, Ross, T. Terlaky, J-Ph Vial, Primal-Dual Algorithms for Linear programming based on the logarithmic barrier method J. of optimization, Theory and Applications, 83: 1-26, 1994. R. Sedgewick, Algorithms in C ++, Addison Wesley, 1992.

W. H. Press et al., Numerical Recipes in C, 2nd ed. Cambridge, 1992. Examples of the operation of the apparatus may be provided where, even in very simple marketing situations, they illustrate the superiority of the matching algorithm remaining at the core of this invention over the conventional "binary equalization" method. These simple cases are described by the transactions that are reassigned by a certain number N of instruments among N traders. This is in the process of an idealization of real market situations, where the exchange rates involved will exist, and the number of items will be different from the number of traders, but even in this case we can explain the general method. In a mathematical language, the reassignment is described as a permutation (an element of the finite group S (N)). From the point of view of the markets, all the permutations are probably equal to represent the movements of the ones provided for the reallocations of the optimal resources, that is, the preferred transactions among the participants of the exchange. However, these transactions can be carried out within a conventional transactional environment (electronic or traditional) being very rare among these permutations. These are the products of change transpositions. In general, the probability of an arbitrarily chosen permutation (potential transaction) is a product of transpositions of changes that converges towards 0 exponentially as N goes to infinity. The main characteristic of the present invention as it applies to this simplified scenario described by the permutations, is that the equalization of the algorithm will allow to satisfy the orders that correspond to any permutation, • not only the permutations of these are in a very special way. This implies that the gain efficiency, as measured by the quotient of the probability of finding an equal (transaction) under the present invention, by comparison with the probability of being made under a conventional group of transactions going to infinity as the number of businessmen and instruments that go to infinity. In conceptual terms, the theoretical group addresses the problem of equalization strongly due to recent developments in the theory of structures, where one of the present inventors has generalized the context which they invented (simple Lie groups) for certain infinite dimensional geometries and geometries. continuous (related to factors 11 (1) in the von Neumann algebra). The theory of structures is significantly interwoven with the theory of the Coxeter groups. And the simple thing about this are the symmetric groups in N letters. The property "Coxeter" in the context of these groups, is expressed by the extraordinary factor where the groups of N letters are essentially amalgamated products of the simple Abelian group of two elements Z / 2Z: the group is generated by Nl transpositions together with the relation where the product of the adjacent transpositions has the order 3: 3: s (I) * 2-1, (s (I) * s (1+ 1)) * 3-l. All the properties of symmetric groups are consequences of these two types of relationships. In addition, all simple Lie groups have finite groups generated by similar relationships, and the complete theory of representations and Lie group classification will not exist without these factors. There are many important combinatorial functions that relate to these groups. The present invention extracts these ideas in the context of equal algorithms. One reason why the present invention is concerned is the fruitful derivation to the fact that the orders and the matches have a discrete (combinatorial) aspect, for example a circular exchange and more general partial exchange classifications, and a continuous one; the size of the orders and the exchange rate under a limit order. Since one of the key ideas in the classification of the theory of the geometric group implicit in the recent search has been the idea of trajectories created between pairs of points that remain in a subset whose structure is determined by the symmetric group, and each stage in the trajectory by a simple transposition, the interest of the present invention assumes that the solution to the equalization problem remains in the paths created from an initial assignment to a final assignment (a partial equalization). In this way, the problem of multidimensional equalization can be related to the problem of traditional binary equalization as the case of the Lie groups of degree generally higher than the group GL (2), where the continuous part of a problem already appears in the grade 1 case where the discrete part of the picture is trivial (Z / 2Z). However, while the language of group theory is relevant to an understanding of the combinatorial type involved in solving the problem of optimal matching, other, more general methods are required to find the solution in the general case.

A solution can not be restricted to find an equalization of the simplified type of the situation only analyzed. A solution must be able to be covered by situations in which there may be several orders to buy a particular resource against other resources, and where the exchange rates between different products may be very different between different orders. Now we must describe a specific simple example in which the present invention can find an optimal match, considering that while the traditional methods can not find an equalization, and at the same time illustrate how the system finds the maximum equalization between possible equalizations, on the basis of of two criteria. We consider the following example of a batch of orders as provided by an N-by-k matrix, where N is the number of orders (in this case), and k is the number of raw instruments (in this case 10) an example of An array of orders is as follows: The interpretation of the matrix is that the coefficient (j, k) represents the amount of the raw resource k-t by the order j-th (the signal indicates where the resource in 'gross is paid or received). A conventional transaction apparatus should consider each group of pairs of raw instruments (45 pairs) as an individual market. The total order flow described by the example should be seen "part by part" in each of the 45 markets. In the example at hand, each of these markets should only see an order, either only a purchase order or only a sales order, and then an appliance should be able to match any of these orders of any magnitude absolutely. The present invention, on the other hand, not only finds an equalization but an optimum equalization, in the sense of the cascade optimization criterion. First you must find an equalization that can maximize the total value (or volume) over all matched orders. Secondarily, I must find, among these, the equalization that can maximize the net profit of the corridor between volume maximization orders, or, in fact, the equalization that can optimize any other appropriate criterion specified. This double optimization can be done either by optimizing the second optimization criterion in a reparameterized space of solutions for the first optimization criterion, or equivalently it can be done within the original space simply by adding the value of the first optimization criterion at an optimal point as a constraint additional to the list of restrictions of the first optimization problem. The general method may work as indicated below in this case.

First, phase 2 of the optimization problem should be defined as indicated below. The variable alpha of the independent vector must be the vector of "degrees of equalization", that is, the vector of degrees to which the individual orders will be accepted (alpha is N-dimensional). In this first phase of optimization, the dependent vol (alpha) function (designated) is the total volume of the total transaction (ie the sum of the absolute values of all the quantities involved in the order). The restrictions for the optimization problem should take the following form: 1. The coefficients (alpha elements) should remain between 0 and 1 (1 must mean full acceptance, 0 not accepted). 2. Each of the lines N of the matrix of orders must be multiplied by the corresponding coefficient of acceptance. The resulting matrix N by k has elements that are functions of the N coefficients. { alpha [j], j = l, ... N} . The constraints then take the following form: for each column of the matrix, the sum column is formed (which again will be a function of alpha [j] 's). Call this function CSj (alpha), j = l, ... k (k the number of raw instruments). Then the restriction j-th is expressed by the condition CSj (alpha) <; 0. This restriction means that the broker (who takes the opposite side of each individual transaction) will hold the non-negative amounts of each instrument if the acceptance is defined by the alpha. The first phase of the optimization is in relation to a fixed exchange rate matrix that is used only for the purpose of calculating the volume function. This exchange rate can be the exchange rate calculated from a matched batch preceded by orders. Orders can, of course, be entered arbitrarily into the exchange rates. Since we have to take the first part of the optimization problem in this way, we can express it implicitly in alpha terms. We will then put a standard technique of restricted optimization to find an optimal solution (with respect to the data described). An optimal solution for the previous example is provided by the vector: (i) alpha = (1, 0, 1, 0 ... 1, 0). Another optimal solution is provided by the vector (ii) alpha = (0, 1, 0, 1 ... 0, 1). In other words, the problem of volume optimization is degenerate. Now we will take a solution to the problem of volume optimization, that is solution (i), to calculate the value of the optimal value of the vol function, and then discard solutions (i) and (ii). Now we will solve the new optimization problem by the variable alpha vector: the return (alpha) is maximum, Where the return is the value of the aggregate transaction of the broker in the exchange rate so that the batch of orders is submitted to the group of restrictions (1 and 2 above) in addition to the additional restriction: vol (alpha) = 20.

Now we will solve this problem by means of one of the standard techniques, and we will find that the unique optimum point for the optimization problem is provided by the alpha point = (0, 1, 0, 1, ..., 0, 1) by which the return of the runner is equal to 20 * epsilon. A further simple example of the operation of the apparatus and method according to the present invention will now be provided. In this simple example, three different raw instruments are marketed among four users (P, Q, R, S,) a particular batch of orders may be stored as in the arrangement indicated below in the first storage medium 18: The first column contains the user code (P, Q, R, S,). Each line represents an example of an order entered by the user in a terminal 12 and transmitted to the server 10 the coefficients in columns 2 to 4 designate the quantities of • the raw instruments that the user wishes to acquire (positive sign) or leave (negative sign). In the example, user P wants to acquire a unit of the raw instrument (I) (second column) (ie a currency or a term in an Inventory Index) for a maximum price of one unit of the raw instrument (II) (third column). The fundamental order matrix denotes that F must be: 1 - . 1 - 1 0 0 1 - 1 - 1 0 1 - 1. 200 . 8 A user has the option of ordering a desired quantity of a first raw instrument at the exchange rate prevailing in the market (an amount that is calculated by a better method of adjusting the previous lots by the system and delivered to the terminal of the user in real time) in terms of a specific amount of the second raw instrument. On the other hand, the user has the option to ignore this type of change that prevails in the market and define their own exchange rate. In the illustrated example, users R and S order the same quantity of a raw instrument (III) for different quantities of the raw instrument (I), which illustrates the fact that the user's orders do not need to be related to the type of change that prevails in the market. If an order was made at the exchange rate prevailing in the market, the CPU 22 can obtain the exchange rate of the third storage medium 26. The system will always calculate the exchange rate that prevails in real time by a better adjustment method using only satisfactory orders, as explained later. Having received the flow and the encoding of the command from the matrix F above, the system now proceeds to produce an optimal match. The equalization produced by the system from the command array may be recorded in the second storage means 20, which contains a MAT series with 4 lines and 3 columns (in the general case the N lines and the k columns where N is the number of orders and k is the number of raw instruments and and N will always have this meaning): The coefficients in the second and third column can be between 0 and 1. This restriction corresponds to the requirements that the user should not have more than one raw instrument that he has ordered, and that he should never receive a negative amount from any raw instrument. organized. The second restriction is that the exchange house (which takes the opposite position of the market) must have non-negative amounts for each raw instrument. This corresponds to the requirements of the exchange house that must not take risks or be exposed, that is to say without low positions in any instrument, without having previous or pending positions that the exchange house may have. The entries in the third column are further restricted by a corresponding one in the second column as discussed above. The first column only represents the order numbers and the system stores the trace from which the user submitted the order numbers. The first MAT column denotes an identification code for the order number in the batch. The second column denotes the degree to which the ordered quantity of the order can be matched (in this case the orders 1, 2 and 4 will be completely matched and order 3 remains totally unmatched) and the third column denotes the degree to which the amount of the raw instrument offered in payment has been accepted in payment. Since order 3 has not been fully supplied, the given order 3 certainly arises not to be paid, henceforth the third line of the third column of the coefficient is 0. In general, there is a restriction where MAT [j, 3] must be less than or equal to MAT [j, 2] since a user can never be forced to pay at a high exchange rate that agrees to pay for his order. When the system operates "naturally", MAT [j, 3] is equal to MAT [j, 2], that is, the system always changes precisely the type of change specified by the user. However, the system can operate from a "deployment control mode" in which the inequality between the second and third columns is possible, and it is by this mechanism that it can be guaranteed that the utility of the runner resulting from the equalization can Do not exceed the given thresholds. The calculations performed by the CPU 22 to optimize the coefficients of columns 2 and 3 of the MAT matrix defined above, which is coded for the degree to which the orders will be accepted. These calculations are carried out by means of a cascade optimization criterion controlled by different apparatuses designated Ap (i). Each device can be a routine or module of the set of programs. The routines must be placed sequentially, Ap (l), ..., Ap (n). The first routine Ap (1) takes the batch of orders F and then returns to the allowable group N by 3 matrices of the coefficient, which satisfies the first optimization criterion, and passes these on the routine Ap (2), and so on until that Ap (n) generates a particular coefficient matrix by designating an optimum coefficient matrix MAT. In the preassigned case, this waterfall will be of a length 2 (but the runner can modify the length). The most important optimization criterion is the liquidity (ie the total value of the transactions executed in the equalization) and the utility of the broker, that is, the spread or difference between what is paid in and what is paid. This is how the broker will earn money within the system. A particular apparatus will now be described in detail with respect to the present example. Ap (l) takes the command array and operates the next transformation on it. We must designate the number of raw instruments for k and the number of orders of the user in the lot by N. First, the matrix D = [B; C] is produced, where B is 8 (in general, 2N) by 4 N dimensional, and C is simply the transposition of the matrix F on (the lines of C corresponding to the indices of the raw instrument, the columns for the user's orders). In general the lines of B will consist of 1 times of the unit vector j-th having a 1 in a column j and in another part the 0's (for lines with an index 2j-l), and -1 times of the unit vector j-th (for lines with an index 2j). In the example, these matrices are the following: The 1-dimesional series b is then produced where b is defined as: b = [1 0 1 0 1 0 1 0 0 0 0] In general, the first 2N entries of b are, alternatively l's and 0's, and the remaining entries are all 0.

A matrix of the exchange rate E is defined and is stored in a third storage medium 26. In this example this is a 1-dimensional series of length k. E - [l, u, v] The entries of E designate the values of unit 1 of the raw instruments I, II and III (in general l, ... k) in terms of the operator's local gross instrument (which is always a raw instrument with the level I, and from here to the first entry of E is always equal to 1. The current data of the exchange rate by themselves are stored in terms of a series of squares such that the entry (j, k) designates the type There is also an apparatus for producing a 1-dimensional series of length N, whose coefficient i-th will be given by the sum of the absolute values of the quantities of the raw instruments comprising the order i-th, multiplied by the relevant coefficient of the matrix E of the exchange rate, we call it the vector OPT, this is a vector of length N. Ap (l) is then an apparatus to perform one of the known algorithms ( for example a polynomial time) returning to a milia of solutions for the problem defined by the optimization problem defined by the volume as an indicated function and by the constraints defined by the two criteria: the 5 orders can be satisfied to a degree between 0 and 1, and the return of the broker must be non-negative on all instruments. More explicitly, this is provided by the following definition (in terms of this example) of 0 the optimization routines involved (the "first optimization routine"): First optimization routine: optimizing-the actual value function of a vector N -dimensional of argument x, defined by:: 5 f (x) = internal product of the OPT series with x and the restrictions are: 0 D.x < that the transposition (b) Where D is (11 by 4 matrix) [B; C], that is to say the matrix whose first line 8 is that of the matrix b (like the previous one) and whose 3 remaining lines are 5 those of C (like the previous ones) .

The optimization apparatus will produce as an output of the group of allowable orders according to the first optimization routine. The shape of this output will be a matrix of the form given in this example as the previous MAT, but with the second column replaced by the transposed vector x expressed as a linear function of parameters, with restrictions on these parameters, and the third column is the group equal to the 2nd.

In the example this is the following group of parameterized matrices that we must designate by o (t): Here the intervals of parameter t are from 0 to 1 The second phase of the optimization then uses an apparatus similar to that described for the optimization of stage 1, to optimize the function G (t) where G (t) = utility of the broker for the order o (t) G (t) is calculated as minus the sum of the terms in each column of the matrix or (t), each one evaluated at the exchange rate of the broker. The second phase of the optimization can be defined explicitly in terms of the apparatus analogously to the construction involving the first stage of optimization as indicated below. First, the second phase of the optimization will involve an apparatus to find the restricted optimum of the following linear function of the 2 * N variables, where the first N coordinates refer, as previously, to the coded coefficients for the degree to which the orders are satisfied, and the second coordinates N for the degree to which the offered payment is accepted. For the purposes of the way described therein, these degrees are the same, but they can not be in the general way. The role of the previous OPT series is now represented by a double-length series 0PT2 whose coefficients are carried out as follows: Firstly, the matrix of orders is divided into two matrices, whose sum is the matrix of the orders, defined singularly by the property that the first matrix has only non-negative coefficients and the second only non-positive coefficients. In addition, for each of these matrices, and for each line, the sum is taken over all the columns, with each entry multiplied by the exchange rate E as above, so that, in the two matrices, positive and negative, it is carried a series of N length. The juxtaposition of these two series provides the series 0PT2. In our example, this series is as indicated below: - In the example under discussion, the associated optimization function of 8 variables is then the sum of the first 3 parameters, plus 0.8 times the fourth parameters, minus the 3 following parameters, minus 1.2 times the last parameter (designating the degree to which the payments involved in the 4 user orders have been accepted). The coefficients in this particular functional form come from the positive and negative components of the F matrix in an obvious way. This is the function that the device seeks to optimize, using the matrix F that emerges from the order flow. Following the constraints of this optimization function are encoded by an apparatus containing the constraints where the coordinates jth and N + jth must be equal (in the general case of the control deployed this restriction must relax), and where the constraints lst N in terms of the parameters lst N are the same as the previous ones, plus the new restriction where the OPT takes the optimal value in these parameters lst N to be the result of the optimization of stage 1. In this way, one observes that the only data of the device involved in the first stage of optimization needs to register the value of the OPT function giving the described restriction.

In the example, the final answer for the cascade optimization routine is the array.

Having been established by means of the optimization phase 1 where everything of P and everything of Q can be satisfied, and where a mixture of R and S (given by a parameter-t-between 0 and 1) is possible as far as the optimization criterion 1 is concerned (the volume at the exchange rate chosen by the operator), the 2nd optimization criterion takes the solution under which the operator receives the highest price for unit 1 of the raw instrument I thus maximizing its usefulness . When the orders are satisfied, they are fed into the exchange rate apparatus, which keeps a storage of new orders satisfied to produce a good adjustment of the exchange rates of the most recent operations. A better example of the adjustment of the algorithm in the case of binary orders is as follows: For each batch of satisfactory orders, the following stages are carried out: 1. Satisfactory orders are added to the storage. Each order defines a margin in the "network of orders" whose node corresponds to a raw instrument on a value date. 2. The group of orders that can be removed without causing the command network to be disconnected is identified again (these are standard algorithms to determine if a network is connected, see for example Sadgewick, chapter 29). If this group is not empty, its old member is removed and stage 2 is repeated. 3. A 1-dimensional series L, whose elements correspond to the vertices of the order graph, is determined to minimize the function DF (L), where DF (L) is the sum over the stored orders of the product of the value of the order and the square of the difference between the logarithm of the exchange rate of the order and the difference between the elements of the L corresponding to the raw instruments of the order (there are standard algorithms to perform at least the latest square optimizations, see for example Press and collaborators, chapter 15). The exchange rate between two pairs of dates of values of the raw instrument can then be calculated as explained d [j] -L [j]), where L [j] and L [j] are the elements of L corresponding to the two pairs of dates of values of the raw instrument, and this is used to update the data of the exchange rate in the storage medium 26. The different conventional systems, offered and offered which have not been satisfied are completely ignored as for what concerns the calculation of the exchange rate. The server tracks the arrival of orders in real time. New received orders are added to the existing batch of incomplete orders. A group of criteria of the lot can be defined by the user, being more important simply the total volume calculated in terms of the local currency of the exchange house, of the orders that comprise the lot. When the criterion is met then the server will calculate an optimal match, using at its core a routine to solve the problems of linear programming, and communicate the result of the equalization (that is, the vector of partial satisfactions) for both the exchange house as to the users. To make sure you get an optimal match in real time, the program will allow you to use a menu of linear programming routines. The program tests each of them for a specified maximum period of time. If the optimization routine ends within time then it returns to its optimal value. The program will then choose the best solution (in lexicographical order with respect to the group of optimization criteria adopted, that is, starting with the main optimization criterion, (which is the volume in the predefined situation) and working its trajectory under the criterion of adopted optimization The best solution is then adopted, and matching is recorded and communicated to the participants.The invention as defined herein involves the principle of "without any market risk." This is true for each individual lot. When a batch of orders has been processed and finds a solution, and it is assumed that the account of the broker and the client's accounts are at zero at the start of the matched order, the position after the orders have been executed will be that the sum of the client's accounts is equal less the account of the broker . The restriction "without any market risk" means that as a result of any batch of matched orders the amounts of each resource involved are transferred from the clients to the broker provided they are non-negative. In other words, the system operator or broker of the clearing house will retain the negative portfolio for the sum of the client's portfolios. The sum of the accounts of the client's portfolio as a result of an equalization of the lot is a non-positive vector in the resource space. This vector is called the deployed vector. Any negative coefficient is strictly called "deployment" won by the runner in the equalization. After the deployment is also inconsiderately defined by any type of change (the equalization is not trusted of any exchange rate matrix, and individual transactions can occur in different exchange rates among themselves). The broker can earn a utility by maintaining the deployment or charging a service charge or a combination of the two. To provide maximum flexibility to the system / corridor operator, a variety of different deployment controls and charge models can be used with the present invention. Two examples are provided below. The first approach involves continuously using the criterion of maximizing the volume analyzed above, and solving the problem of deployment control by the discount method of over-deployment. The discount is defined by a parameter rb between 0 and 1 that indicates the degree to which the deployment will be reimbursed: 1 means that the deployment is reimbursed to the maximum possible extent and 0 is not reimbursed for the whole. The exchange rate matrix E, which remains in this algorithm, can be calculated in the manner already indicated above or in any other appropriate manner. Note that in this approach, customers whose orders placed under the prevailing exchange rate and which are matched are not penalized in any way for their "bad retribution". Once the exchange rate is defined, the value of the accepted orders can be calculated with the exchange rate. Some of these orders may have positive value and others may be of negative value. The orders (accepted portions of) with negative value "overpaid orders" will be named. The deployment control is a mechanism for a redistribution portion rb of the total deployment for the traders whose overpaid orders have been accepted. To do this, first calculate the overpayment coefficient of an overpaid order as the value of the overpayment divided by the total value of all overpaid orders. These coefficients represent a unit partition. The deployment control mechanism is then implemented to accredit user accounts that correspond to all orders overpaid to an account that at the exchange rate used for this calculation, is equal to the deployment times for the coefficient of the times rb of the overpayment. Credit for overpaid orders will be distributed in sequence to users in decreasing order according to their overpayment coefficient. The broker will pay for the deployment in the currency in which the deployment was sustained, starting with the highest retention (in value) and low as far as it is necessary for the credit of the clients with respect to their overpaid orders. The discount for each order is subtracted from the amount to be paid with respect to the order for the exchange house before the amounts involved in the match are placed. In short, the discount in question is worked as indicated below: An optimal match is as before. The exchange rate matrix E is calculated.

The discount is calculated in the above manner using this exchange rate, and the amounts payable to and by the exchange house are adjusted according to the discount. A second approach, referred to as the central equalization of the exchange rate, will now be described. This is more radical because it modifies the matching algorithm described above sacrificing the principle of the maximum volume obtained for a given order flow. A control parameter of the theta deployment is specified > 0. These limits (or optionally determine) the utility of the broker, as a fraction of the total volume traded. Each batch is then processed as follows: 1. A central exchange rate is determined for the batch. There are several possible ways to do this; two will be described below. 2. A batch derived from orders is created as follows: Each order is processed according to its "generosity" g. determined as the proportion of the exchange rate of the order for the corresponding exchange rate. For this purpose the exchange rate of an order has been defined to be the reciprocal part of the amount of the pair of value dates of the gross instrument ordered for the quantity of the pair of value dates of the gross instrument offered in the exchange. There are three cases: i) g > + teta: the order is added to the new lot, with its exchange rate adjusted so that g = l + teta. ii) 1 < g < 1 + teta: the order can optionally be included in its original exchange rate, in the case that it is not desired to give strength to a minimum payment of the deployment. iii) g < 1: the order is excluded from the new batch. 3. The derived lot is then matched as is normally done, and the satisfaction coefficient of each order is updated according to the quantity traded of a designated principal component of the order. A simple technique for determining the fixed exchange rate performs a non-deployed controlled equalization in the entire lot, and then calculates the exchange rate in the usual way using the result of this atrial matching. A preferred, though computationally more expensive, determination of the fixed exchange rate by which the previous algorithm addresses the largest volume marketed. This requires the use of a non-linear optimization routine (see for example, Press et al., Chapter 10). In this case, the indicated function for optimization can be calculated by performing the same tests for the assumed fixed exchange rates. The above algorithm is easily generalized to include orders that contain more than two components. In this case the derived orders are formed by the projection in a "theta hyperplane" for the space, which consists of the orders that are more generous than the fixed exchange rate by a factor of 1 + teta, while those whose type of The specified change is less generous than the excluded ones. This projection can take place along either the negative or positive components of the order, according to which the user is willing to allow the discrepancy. In the first discount in question analyzed above, all the orders that are satisfied under the original matching algorithm are still satisfactory. This is not any real time under the simple exchange rate algorithm. Different conventional systems of change, which are based on the construction of direct exchanges between the counterparts, the present invention allows the user to enter a wide range of different order dimensions. This is due to the fact that, from the mathematical point of view, the optimization problem is also defined and in fact does not vary under the operations of fractions to a very large order in several small ones, or adding any finite number of small orders in one big order In addition, there are effective numerical optimization methods that can practically implement the solution to the optimization problem. As the structure of the most natural optimization problems that arise (as listed above) is linear, any of the numerical schemes will be used to solve linear optimization problems. The present invention has the advantage of enabling transactions that are genuinely non-binary: they are not exchanged between the two parties, but are "non-binary exchanges" of items of k between n merchants. An equalization can occur between 3 users, P, Q and R, even if P and Q, Q and R, and P and R, are isolated pairs of users who can not be in a position to trade. For example, if P, Q, R hold, respectively, unit 1 of three items p, q, r (stocks, options or money), but want to buy (respectively) q, r, p, at exchange rates all equal to 1, then the system can automatically produce an equalization by assigning qa P, ra Q, and p R. Similar examples can be constructed that illustrate the overall capacity of the system to build matches that take the direct advantage of the full range of market share, they are concerned with respect to the entire group of users and the entire group of instruments. Assuming, for the reason of the argument, that there are at any given time users within the system located to, involve 500,000 financial instruments. It is possible to build examples that show that an equalization is possible, and may in fact be constructed by the algorithm, where there will be no satisfactory orders, that is, no possible deal among the 499,999 users who market 499,999 different financial instruments. In other words, the equalization algorithm allows global transactions between a set of users in a dimensionally larger instrument market that can not be carried out if any sub-group of users makes agreements between themselves. Equalization in accordance with these modalities is not based on agreements between individuals but on a numerical procedure by which an equalization, that is, an allocation of resources among the negotiators, is implemented, and this numerical procedure is centrally calculated, without reaching Raising awareness and then the specifically informed process involves individual traders. For this reason it is not necessary for the particular users who are aware of the other orders in the system. In this way anonymity can be preserved, which is a key advantage for users who can place orders without fearing that their entry into the market will disturb the stock market in a direction opposite to their own interests. However, a user also has the option of making certain orders visible to other participants in the exchange. These orders can contribute to complete market information in the following way. An additional mode of operation of the present invention is to allow hypothetical orders to undergo matching against the orders marked as visible. Such orders can be used to obtain complete market information. The orders are hypothetical in that the orders and their respective coefficients of satisfaction are not sent to the party for the payment, on the other hand the coefficients or the information derived from the coefficients are returned to the user. Two examples of this to obtain complete market information are indicated below: (a) a user wants to know what particular dimensions of the exchange rate of the orders he can satisfy, for example, the user wants to know the exchange rates where 10, 20 or 30 million US dollars can be purchased in yen. To do this the system repeatedly submits the hypothetical orders for 30 million US dollars (the largest of the amounts that the user wants to have the complete market information approximately) in a group of prices, that is 120 yens / per US dollar ( very low exchange rate) 121 yes / per US dollar and so on up to 140 yen / per US dollar (very generous exchange rates). The system then returns the respective coefficients, that is, the quantities of these orders that can be matched. For example, in the given example, the coefficients can be 0.0,0.2,0.4,0.6,0.8, 1.0, 1.0 ... which for an order of 30 million US dollars corresponds to the quantities of the order that can be satisfied : 0 to 120 yens / per US dollar, 0 to 121 yen / per US dollar, 6 million dollars at 122 yen / per US dollar, 12 million dollars at 123 yen / per US dollar, 18 million dollars at 123 yen / per US dollar, 24 million dollars at 125 yen / per US dollar, 30 million dollars at 126 yen / per US dollar and so on. In this case the user may be informed that an order for 10 million US dollars must be paid at an exchange rate of 123 yen / US dollar, an order of 20 million US dollars must be paid at an exchange rate of 125 yen / per US dollar and an order for 30 million US dollars must be satisfied at an exchange rate of 126 yen / US dollar. (b) alternatively, the user may wish to know the largest quantities of a resource that can be purchased at a particular exchange rate, for example a user should want to know as some US dollars at the exchange rates of 130, 131, and 132 yen / per US dollar. In this case the reiterative system will submit the high-volume hypothetical orders (in practice equal to the largest volume recorded by any order submitted to the system in this particular class of the instrument / order pair), for example two hundred million dollars, in each of the prices submitted by the user. These orders are processed in the usual way and the partial satisfaction coefficients are returned. In the previous example these should be 0.25, 0.3, and 0.6. The user can then inform himself that orders of 50 million US dollars must be paid at 130 yen / US dollar, 60 million US dollars at 131 yen / US dollar and 120 million US dollars at 132 yen / US dollar ( the dimension of the orders is provided by the product of the coefficient and the hypothetical order of 200 million dollars where it was submitted). Although some of the previous examples have concerned financial instruments, this is not meant to limit the scope of protection defined by the claims. For example, the invention is applicable to systems for assigning calculation time, telecommunication and broadband frequencies, power generation and distribution capacity and so on. The invention can be implemented by means of a computer and the invention therefore includes a computer readable storage medium that has registered therein a program that contains the code of the components that when loaded on a computer will cause the computer to operate according to the method of the invention.

Glossary The following definitions are provided to assist in understanding the terminology used in the examples described above, but are not limiting of the present invention. Raw instrument: currencies, equities, etc., with no time component. Simple Resource: a quantity of a raw instrument. Portfolio Space (PF): the vector space with base indexed by the group of raw instruments. PF (+): the group of position vectors in PF with non-negative components. Simple resources are particular elements of this group. Resource space - finite subsets of PF (+) x T with any of the two elements that have the different time component (T). The T is the positive time axis and the T coordinate of a point in this space is known as the value date. PF (+) is identified with the subset consisting of a single element of the form (pf, now). Compound resources: Points in the resource space. Resource flow: a group of compound resource pairs, as specified by the order. Orders are of the form: for the purchase of a compound resource against another compound resource. Simple derivative: a flow of resources (ie, an order) with the rule that, if the order is satisfied, the negotiator will be required to receive or release (according to whether the value of the quantity is positive or negative) the specified amount of each simple resource on its specified value date. Option: an instrument that provides an appropriate market participant (but without obligations) to exchange resources at some date in installments (as opposed to the simple derivatives in which the installment transfers of simple resources are obligatory in all the parties).

It is noted that in relation to this date, the best method known to the applicant to carry out the aforementioned invention, is the conventional one for the manufacture of the objects or products to which it refers.

Claims (58)

1. Having described the invention as above, the content of the following claims is claimed as property: 1. An apparatus for processing commercial orders, the apparatus comprises a central server connectable to a plurality of terminals in which the user's orders are inserted , characterized in that the central server comprises: communication means for receiving the user's commands from the terminals through a network; a first storage means for storing the received user commands as a series whose elements define a particular first resource ordered by a particular user; a second storage means for storing a series of coefficients each representing the proportion of a particular order that is satisfied; and a processing means for recovering the orders of the first storage means, calculating an optimized group of values of the coefficients with respect to at least one predetermined adjustable constraint, and at least one adjustable, predetermined, and stored criterion of the values of the coefficient optimized in the second storage means, the communication means are also for transmitting the processed orders and their respective coefficients.
2. 2. An apparatus according to claim 1, characterized in that at least one restriction includes that the value of each of the coefficients. Is less than or equal to 1 and greater than or equal to 0.
3. 3. An apparatus according to any of claims 1 to 2, characterized in that the processing means is adapted to process the commands such that a designated user takes the opposite position to each of the other commands of the user agreeing to change a proportion of the first resource ordered for a second resource, where the proportion corresponds to the optimized coefficient for the order.
4. 4. An apparatus according to claim 3, characterized in that at least one restriction includes that if all the orders were completed, in proportion to their respective coefficients, the designated user's retentions arise from the processed orders which will be only negative quantities of each resource, including after the maturation of all derivatives and simple options for resources marketed in the future.
5. 5. An apparatus according to any of claims 3 or 4, characterized in that at least one criterion includes maximizing the utility of the designated user, in terms of a particular simple resource, based on an exchange rate.
6. 6. An apparatus according to any of the preceding claims, characterized in that the central server additionally comprises a third storage means for storing a series of data representing the current exchange rate between each resource and at least another resource, and wherein the processing means is also for recovering the data of the third storage means.
7. 7. An apparatus according to any of the preceding claims, characterized in that at least one criterion includes maximizing the volume given by the sum of the absolute values of the components of all the orders that are satisfied, partially or completely, in terms of a resource particular simple at a given exchange rate.
8. 8. An apparatus according to any of the preceding claims, characterized in that the processing means is adapted to optimize the values of the coefficients by successively applying the respective criteria in a cascade manner.
9. 9. An apparatus according to claim 6, characterized in that it additionally comprises means for specifying the cascade criterion sequence.
10. 10. An apparatus according to any of the preceding claims, characterized in that the processing means is adapted to apply, in sequence, each of the plurality of predefined linear programming routines, or convex programming routines, or standard combinatorial optimization techniques. , to optimize the coefficients until one of the following events occurs: a specified maximum period of time elapses. an optimal solution is found.
11. 11. An apparatus according to claim 10, characterized in that if a specified maximum period of time elapses before a solution is found, a consistent suboptimal solution is used to obtain the optimized group of the coefficient values.
12. 12. An apparatus according to any of the preceding claims, characterized in that the processing means is adapted to optimize the coefficients for the batches of the orders received.
13. 13. An apparatus according to claim 12, characterized in that the processing means is adapted to determine the end of a batch by a present interval of time that has elapsed since the start of the batch.
14. 14. An apparatus according to claim 12, characterized in that the processing means is adapted to determine the end of a batch by the total value of the order exceeding a threshold value.
15. 15. An apparatus according to any of claims 12 to 14, characterized in that it is adapted to carry out unsatisfactory orders, completely or partially, in a batch for the next batch.
16. 16. An apparatus according to any of claims 12 to 15, characterized in that the orders are unsatisfied, completely or partially, after a length of time of presentation of these orders is adapted to remove from the first storage means.
17. 17. An apparatus according to claim 16, characterized in that the predetermined duration of time for each order is specified by the interested user.
18. 18. An apparatus according to any of the preceding claims, characterized in that unsatisfactory commands are removed from the first memory means in a user's request.
19. 19. An apparatus according to any of the preceding claims, characterized in that at least one user order stored in the first storage means specifies a second particular resource offered in the change for the first resource to define a resource flow.
20. 20. An apparatus according to any of the preceding claims, characterized in that at least one order from the user stored in the first storage means orders the first resource at the exchange rate prevailing in the market.
21. 21. An apparatus according to any of the preceding claims, characterized in that a resource in at least one order is a composite resource.
22. 22. An apparatus according to any of the preceding claims, characterized in that at least one of the terminals is connected to the central server by means of a sub-server that adds the user's commands.
23. 23. An apparatus according to any of the preceding claims, characterized in that the communication means is adapted to transmit the commands using TCP / IP.
24. 24. An apparatus according to claim 6 or "with any dependent claim of claim 6, characterized in that the processing means calculates and updates the exchange rates in the third storage means based on the flow of successful orders.
25. 25. An apparatus according to any of the preceding claims, characterized in that the instruments sold are financial, such as currencies, guarantees, and terms on the value of the articles.
26. 26. An apparatus according to any of the preceding claims, characterized in that the communication means transmits the processed commands and their coefficients to an additional apparatus for the liquidation of the orders.
27. 27. 'A computer terminal characterized in that it comprises: a means of communication for receiving the processed orders and their respective coefficients of an apparatus according to any one of claims 1 to 25; and a device for activating the transfer of resources according to the filled part of each order specified by the respective coefficient.
28. 28. A method for processing commercial orders characterized in that it comprises the steps of: receiving from the users the orders each specifying a first particular resource ordered by a particular user and stored in series in a first storage means; processing the orders retrieved from the first storage means to calculate a group of coefficients each representing the proportion of a particular order that is satisfied; optimizing the values of the coefficients with respect to at least one adjustable, predetermined constraint and to at least one adjustable, predetermined criterion; generate the processed orders and their respective coefficients.
29. 29. A method according to claim 28, characterized in that at least one restriction includes that the value of each coefficient is at least less than or equal to 1 and greater than or equal to 0.
30. 30. A method according to claim 28 or 29, characterized in that a user designated user takes the opposite position to each of the other orders of the user agreeing to change a proportion of the first resource ordered for a second resource, where the proportion corresponds to the optimized coefficient for the order.
31. 31. A method according to claim 30, characterized in that at least one restriction includes that if all the orders were completed, in proportion to their respective coefficients, the retentions of the designated user will arise from the processed orders which should be only non-negative quantities of each resource, including after the maturation of all options and simple derivatives for resources marketed in the future.
32. 32. A method according to claim 30 or 31, characterized in that the optimization step includes a criterion that maximizes the utility of the designated user, in terms of a simple particular resource, based on an exchange rate.
33. 33. A method according to any of claims 28 to 32, characterized in that a third storage means is for storing a series of data representing the current exchange rate between each resource and at least one other resource, the additional method comprises the stage of recovering the date of the exchange rate from a. third storage medium to be used in the optimization of the coefficients.
34. 34. A method according to any of claims 28 to 33, characterized in that the step of optimization includes maximizing the volume given by the sum of absolute values of the components of all the orders that are satisfied, partially or by all the orders that are satisfied , partially or completely, in terms of a simple particular resource at a given exchange rate.
35. 35. U - method according to any of claims 28 to 34, characterized in that the optimization step further comprises successively applying respective criteria in a cascade fashion to obtain optimized values of the coefficient.
36. 36. A method in accordance with the claim 35, characterized in that it additionally comprises the step of specifying the sequence of the criteria that fall in the form of a cascade.
37. 37. A method according to any of claims 28 to 36, characterized in that the optimization step further comprises applying, in sequence, each of the plurality of predefined linear programming routines, or the convex programming routines, or standard optimization techniques. combinatorial, to optimize the coefficients until one of the following events occurs: a specified maximum period of time elapses. an optimal solution is found.
38. 38. A method according to claim 37, characterized in that if a specified maximum period of time elapses before an optimal solution is found, a consistent suboptimal solution is used as the optimized group of coefficient values.
39. 39. A method according to any of claims 28 to 38, characterized in that the processing step additionally comprises recovering the orders of the second storage medium of the batches, and followed by the optimizing step to obtain the optimized coefficient values for the batches of the orders.
40. 40. A method according to claim 39, characterized in that the end of a batch is determined by a time interval present from the beginning of the batch.
41. 41. A method according to claim 39, characterized in that the end of the batch is determined by the total value of the order exceeding a threshold value.
42. 42. A method according to any of claims 39 to 41, characterized in that it additionally comprises the step of carrying out the orders in a batch that are not satisfied, completely or partially, following the optimization stage, to be processed in the next batch.
43. 43. A method according to any of claims 39 to 42, characterized in that it additionally comprises the step of withdrawing the orders from the second storage means that have not been fulfilled, completely or partially, after a predetermined period of time of presentation of the orders.
44. 44. A method according to claim 43, characterized in that the predetermined time period for each order is specified by the interested user.
45. 45. A method according to any of claims 28 to 43, characterized in that it additionally comprises the step of erasing the unsuccessful commands in the user's request from the second memory means.
46. 46. A method according to any of claims 28 to 45, characterized in that at least one user command stored in the first storage means specifies a second resource offered in the change for the first resource to define a flow of resources.
47. 47. A method according to any of claims 28 to 46, characterized in that at least one user order stored in the first storage means orders the first resource at a prevailing market exchange rate.
48. 48. A method according to any of claims 28 to 47, characterized in that a resource in at least one order is a composite resource.
49. 49. A method according to any of claims 28 to 48, characterized in that it additionally comprises the step of communicating the commands introduced in a plurality of terminals to a central server for the processing of orders, by means of a network.
50. 50. A method according to claim 49, characterized in that it additionally comprises the steps of adding the orders of the users in a subserver before communicating them to the central server.
51. 51. A method according to claim 49 or 50, characterized in that the communication is carried out by means of TCP / IP.
52. 52. A method according to any of claims 28 to 51, characterized in that it additionally comprises the step of calculating the updated exchange rates based on the flow of satisfactory orders and storing the updated exchange rates in the third storage means.
53. 53. A method according to any of claims 28 to 52, characterized in that the instruments sold are financial, such as currencies, guarantees, and terms of the values of articles.
54. 54. A method according to any of claims 28 to 53, characterized in that it additionally comprises the step of transmitting the result of the generation stage to a means for the liquidation of the orders.
55. 55. A method according to any of claims 28 to 54, characterized in that a proportion of the value of an accepted order greater than a prevailing exchange rate is reimbursed to the respective user.
56. 56. A method according to any of claims 28 to 54, characterized in that a designated user receives a utility limited by, or predetermined as, a fraction of the total volume traded.
57. 57. A method according to any of claims 28 to 56, characterized in that it comprises the step of controlling a process using the processed orders and their coefficients generated in the generation stage.
58. 58. A computer-readable storage medium that has registered therein a program that contains the component codes that, when loaded and executed, will cause the computer to operate according to the method of any of the claims preceded by the method. APPARATUS AND METHOD OF PROCESSING OF ORDERS SUMMARY OF THE INVENTION An apparatus for processing commercial transactions, comprising: a central server connected to a plurality of terminals in which the user's orders are entered. The central server additionally comprises: a communication means for transmitting the user's commands between the terminals and the central server by means of a network; a first storage means for storing as received the user's orders whose elements define a particular first resource ordered by a particular user; the first storage medium is also used to store the resources that can be offered by the users for the change against the first resource ordinao; a second storage means for storing a series of coefficients each representing the proportion of a particular sentence that is satisfied; a processing means for recovering the sentences to the first storage means, calculating an optimized group of values of the coefficients with respect to at least one adjustable constraint, predetermined and at least one adjustable, predetermined criterion, storing the values of the coefficient optimized in the second storage medium; and output means for communicating the processed orders and their respective coefficients.