CN115829169B - Business processing method and device based on mixed integer linear programming - Google Patents
Business processing method and device based on mixed integer linear programming Download PDFInfo
- Publication number
- CN115829169B CN115829169B CN202310126641.0A CN202310126641A CN115829169B CN 115829169 B CN115829169 B CN 115829169B CN 202310126641 A CN202310126641 A CN 202310126641A CN 115829169 B CN115829169 B CN 115829169B
- Authority
- CN
- China
- Prior art keywords
- information
- service
- target
- solving
- business
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the specification provides a service processing method and device based on mixed integer linear programming, wherein the service processing method based on mixed integer linear programming comprises the following steps: receiving a service prediction request aiming at a target service, wherein the service prediction request comprises service expected information and service constraint information corresponding to the target service; decomposing the business expected information into a plurality of business solution sub-information based on the business constraint information; and respectively solving the plurality of service solving sub-information to obtain a target prediction sampling value corresponding to the target service. According to the method, the mixed integer linear programming problem is decomposed into the plurality of service solving sub-information according to the service constraint information, the service solving sub-information are mutually independent, the parallel calculation can accelerate the solving speed of the service solving sub-information, and the subsequent calculation efficiency is improved.
Description
Technical Field
The embodiment of the specification relates to the technical field of computers, in particular to a business processing method based on mixed integer linear programming.
Background
In many business scenarios, a solution problem of a mixed integer linear programming model is encountered, such as energy power, electronic commerce, a supply chain, cloud computing, chemical industry, finance, education and scientific research, for example, energy power is taken as an example, each stage of power grid needs to determine a unit combination in a future period of time, a scheme is formulated for meeting the power consumption requirement of a user, when each generator unit is started and shut down in the future period of time and outputs power at each time point is specified, both the physical requirement is met, and the power generation cost is as low as possible.
As service scenarios become more and more abundant, the service scale of a single service scenario becomes larger and larger, and the traditional branch-and-bound method needs longer time to obtain a satisfactory solution, but in an actual service scenario, the time for solving a mixed integer linear programming task model for each service is limited, and the traditional method is more and more difficult to adapt to a new situation, so that a more rapid way is needed to obtain an optimal solution of a target service.
Disclosure of Invention
In view of this, the embodiments of the present specification provide a business processing method based on mixed integer linear programming. One or more embodiments of the present specification relate to a business processing apparatus based on mixed integer linear programming, a computing device, a computer readable storage medium, and a computer program to solve the technical drawbacks of the prior art.
According to a first aspect of embodiments of the present disclosure, there is provided a service processing method based on mixed integer linear programming, including:
receiving a service prediction request aiming at a target service, wherein the service prediction request comprises service expected information and service constraint information corresponding to the target service;
decomposing the business expected information into a plurality of business solution sub-information based on the business constraint information;
and respectively solving the plurality of service solving sub-information to obtain a target prediction sampling value corresponding to the target service.
According to a second aspect of embodiments of the present disclosure, there is provided a unit combination prediction method based on mixed integer linear programming, including:
receiving a service prediction request for a unit combination prediction service, wherein the service prediction request comprises service expected information and service constraint information corresponding to the unit combination prediction service;
decomposing the business expected information into a plurality of business solution sub-information based on the business constraint information;
and respectively solving the plurality of service solving sub-information to obtain the prediction unit combination corresponding to the unit combination prediction service.
According to a third aspect of embodiments of the present specification, there is provided a service processing system based on mixed integer linear programming, comprising
The terminal side equipment is used for generating a service prediction request aiming at a target service and sending the service prediction request to the cloud testing equipment, wherein the service prediction request comprises service expected information and service constraint information corresponding to the target service;
the cloud testing equipment is used for receiving the service prediction request, decomposing the service expected information into a plurality of service solving sub-information based on the service constraint information, respectively solving the plurality of service solving sub-information to obtain a target prediction sampling value corresponding to the target service, and sending the target prediction sampling value to the terminal side equipment.
According to a fourth aspect of embodiments of the present specification, there is provided a service processing apparatus based on mixed integer linear programming, comprising:
the system comprises a receiving module, a service prediction module and a service constraint module, wherein the receiving module is configured to receive a service prediction request for a target service, and the service prediction request comprises service expected information and service constraint information corresponding to the target service;
a generation module configured to decompose the business desire information into a plurality of business solution sub-information based on the business constraint information;
and the prediction module is configured to solve the plurality of service solving sub-information respectively to obtain a target prediction sampling value corresponding to the target service.
According to a fifth aspect of embodiments of the present specification, there is provided a computing device comprising:
a memory and a processor;
the memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions, where the computer executable instructions when executed by the processor implement the steps of the above-described business processing method based on mixed integer linear programming or the set combination prediction method based on mixed integer linear programming.
According to a sixth aspect of embodiments of the present specification, there is provided a computer readable storage medium storing computer executable instructions which, when executed by a processor, implement the steps of the above-described mixed integer linear programming based business processing method or mixed integer linear programming based crew combination prediction method.
According to a seventh aspect of embodiments of the present specification, there is provided a computer program, wherein the computer program, when executed in a computer, causes the computer to perform the steps of the above-described business processing method based on mixed integer linear programming or the set combination prediction method based on mixed integer linear programming.
According to the method and the device, the mixed integer linear programming problem is decomposed into the plurality of business solving sub-information according to the business constraint information, the plurality of business solving sub-information is processed in parallel, the solving result corresponding to each business solving sub-information is obtained, the business solving sub-information is independent, the parallel calculation can accelerate the solving speed of the business solving sub-information, and therefore the solving speed of the mixed integer linear programming problem is improved, finally, the target prediction sampling value of the mixed integer linear programming problem is obtained according to the result of each business solving sub-information, the optimal solution can be rapidly screened out in the limited solving result, the solving speed is further accelerated, and the calculating time is saved.
Drawings
Fig. 1 is a schematic diagram of a service processing method based on mixed integer linear programming according to an embodiment of the present disclosure;
fig. 2 is a flowchart of a service processing method based on mixed integer linear programming according to an embodiment of the present disclosure;
FIG. 3 is a schematic diagram of a sampling feasible region provided by one embodiment of the present description;
FIG. 4 is a schematic diagram of determining an initial sample value provided by one embodiment of the present disclosure;
FIG. 5 is a process flow diagram of a business processing method based on mixed integer linear programming applied to a dock-allocated berth scenario provided in one embodiment of the present disclosure;
FIG. 6 is a flow chart of a method for unit combination prediction based on mixed integer linear programming according to one embodiment of the present disclosure;
fig. 7 is a schematic structural diagram of a service processing device based on mixed integer linear programming according to an embodiment of the present disclosure;
FIG. 8 is a schematic diagram of an apparatus for a mixed integer linear programming based crew combination provided in one embodiment of the present disclosure;
FIG. 9 is a hybrid integer linear programming based business processing system provided in one embodiment of the present disclosure;
FIG. 10 is a block diagram of a computing device provided in one embodiment of the present description.
Detailed Description
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present description. This description may be embodied in many other forms than described herein and similarly generalized by those skilled in the art to whom this disclosure pertains without departing from the spirit of the disclosure and, therefore, this disclosure is not limited by the specific implementations disclosed below.
The terminology used in the one or more embodiments of the specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of the specification. As used in this specification, one or more embodiments and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used in one or more embodiments of the present specification refers to and encompasses any or all possible combinations of one or more of the associated listed items.
It should be understood that, although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, a first may also be referred to as a second, and similarly, a second may also be referred to as a first, without departing from the scope of one or more embodiments of the present description. The word "if" as used herein may be interpreted as "at … …" or "at … …" or "responsive to a determination", depending on the context.
It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) referred to in this specification are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data are required to comply with the related laws and regulations and standards of the related country and region, and are provided with corresponding operation entries for the user to select authorization or rejection.
First, terms related to one or more embodiments of the present specification will be explained.
And (3) unit combination: a solution is formulated in the power grid to meet the electricity demand of users, so as to define when each generator set is started and shut down and outputs power at each time point in a future period, the solution needs to meet a series of physical requirements, and the total power generation cost needs to be as small as possible, and the problem is generally modeled as a mixed integer linear programming, especially a mixed 0-1 linear programming.
Mixed integer linear programming: the method is that variables (all or part) in planning are limited to integers, if the variables are limited to integers in a linear model, the linear model is a mixed integer linear programming, and the method for solving the mixed integer linear programming is only suitable for the mixed integer linear programming.
Hybrid 0-1 linear programming: the method is a mixed integer linear programming, and the objective function and the constraint condition are both linear functions of decision variables, wherein a part of the decision variables can only take the value of 0 or 1.
The method comprises the following steps: refers to solutions that can satisfy all constraint conditions in the linear programming problem.
Linear programming relaxation: the constraint that requires integer numbers in mixed integer linear programming is canceled, and a linear programming model, namely linear programming relaxation (LinearProgramming Relaxation, LPR), can be obtained, and because the linear programming model can be generally solved quickly, a feasible solution is generally constructed based on the solution of the LPR when solving the mixed integer linear programming problem.
In many business scenarios, a solution problem is typically encountered for a mixed integer linear programming model, including objective functions and constraints. Taking the unit combination of the power grid scene as an example, each stage of power grid needs to determine the unit combination of the next day every day, usually the problem is converted into a mixed 0-1 linear programming model, and a branch-and-bound method is used for solving the corresponding mixed 0-1 linear programming model to obtain an optimal solution or near-optimal solution.
As the number of users and units of the power grid increases, the corresponding model scale increases, and the traditional branch-and-bound method needs longer time to obtain a satisfactory solution, but the time reserved for solving the model in the actual production of the power grid is limited, so that the traditional method is more and more difficult to adapt to a new situation.
In the conventional branch-and-bound method, integer constraint of a model is usually ignored first, corresponding variables are regarded as common continuous variables, so as to obtain a Linear Programming (LP) model, so as to be called as Linear Programming Relaxation (LPR) of an original problem, a solution of the relaxation problem can be obtained by using the LP model to solve, so as to be called as relaxation solution, a corresponding objective function is recorded as a lower bound, an objective function value of the original problem cannot be smaller than the lower bound, and if values of corresponding variables in the relaxation solution exactly meet integer requirements of the original problem, the relaxation solution is an optimal solution of the original problem, so that the solution is ended.
If the corresponding variable in the relaxation solution does not meet the integer requirement of the original problem, branching is carried out, namely the variable is respectively fixed to two LP sub-problems with 0 and 1 structures, the two LP sub-problems are respectively solved, if the obtained solution meets the integer requirement, a feasible solution of the original problem is obtained, and a corresponding objective function value is recorded as an upper bound, namely the target value of the original problem cannot be larger than the upper bound; if the obtained solution still does not meet the integer requirement, continuing to carry out branch solution on the LP sub-problem, updating the upper bound and the lower bound, and obtaining the optimal solution when the upper bound and the lower bound are combined.
The conventional branch-and-bound method is a tree search method that requires traversing each leaf node from the root node, where upper and lower bounds are used to prune the worse branches to reduce the search space, and in the worst case, the method requires traversing all leaf nodes. Because the solution of the leaf node depends on the father node, when the node of the feasible solution is deeper, a series of sub-problems need to be sequentially solved to achieve the feasible solution of the node, and the parallel calculation cannot be utilized to accelerate the sequential process, so that in practical application, the speed of solving the problem by the method cannot be effectively improved by using large-scale parallel calculation.
The academic world also provides a method for solving based on random sampling in the hypercube, and the method can solve the corresponding sub-problems in a parallel mode because the acquired points are independent of each other, but the method can acquire a large number of invalid points which lead to infeasible solutions, so that the efficiency is quite low and the method cannot be practically applied.
Based on this, in the present specification, a service processing method based on mixed integer linear programming is provided, and the present specification relates to a unit combination prediction method based on mixed integer linear programming, a service processing apparatus based on mixed integer linear programming, a service processing system based on mixed integer linear programming, a computing device, and a computer readable storage medium, which are described in detail in the following embodiments one by one.
Referring to fig. 1, fig. 1 is a schematic diagram illustrating a service processing method based on mixed integer linear programming according to an embodiment of the present disclosure, and as shown in fig. 1, the service processing method based on mixed integer linear programming provided in the embodiment of the present disclosure is applied to a terminal 100, where the terminal 100 may be a terminal device such as a notebook computer, an intelligent terminal, a server, a cloud server, or the like.
In the terminal 100, a target business original mixed integer linear programming model corresponding to the target business is created according to the target business, and the target business original mixed integer linear programming model generally includes business expectation information and business constraint information, where the business expectation information specifically refers to an objective function corresponding to the target business, for example: min 3x 1 +5x 2 -2x 3 . Obtaining the minimum objective function according to the objective task expectationx 1 、x 2 、x 3 . Wherein, the liquid crystal display device comprises a liquid crystal display device,x 1 、x 2 、x 3 in particular parameters in the target task.
The traffic constraint information refers in particular to the condition that several parameters in the objective function need to satisfy, e.g. 5x 1 +x 3 ≥3，3x 2 -3.5x 3 ≤1，x 1 ≥0，x 2 ≥0，x 3 E {0,1}. In particularThe traffic constraint information generally includes first traffic constraint information and second traffic constraint information, wherein the first traffic constraint information includes integer decision parameters, i.e. the first traffic constraint information refers to limiting the type of parameters, specifically, the parameters in the objective function are integers, for example x 3 E {0,1} representsx 3 Only values in 0 or 1 can be taken. The constraint condition is first service constraint information, and the second service constraint information specifically refers to a linear constraint condition in the service constraint information, for example, 5x 1 +x 3 ≥3，3x 2 -3.5x 3 ≤1，x 1 ≥0，x 2 ≥0。
Then, the first business constraint information in the target business original mixed integer linear programming model is subjected to linear programming relaxation to generate a target business linear programming relaxation model, specifically, parameter constraint requiring integer in the first business constraint information is cancelled and adjusted to be linear constraint, for example, the first business constraint information is subjected to linear constraintx 3 E {0,1} is adjusted to be 0.ltoreq.x 3 ≤1。
After the target business linear programming relaxation model is obtained, a plurality of target sampling values are generated based on constraint conditions of the target business linear programming relaxation model, wherein the target sampling values specifically refer to possible values of parameters meeting first business constraint information and second business constraint information, the target sampling values are respectively transmitted into the target business linear programming relaxation model, initial solution information corresponding to each target sampling value is calculated respectively, one target solution information is selected from a plurality of initial solution information, the target sampling value corresponding to the target solution information is determined to be a final target prediction sampling value, a specific selection process is related to a target function, for example, the target function is to find the minimum value, and the minimum initial solution information is selected from the plurality of initial solution information to be the target solution information; if the objective function is to find the maximum value, the largest initial solution information is selected from the plurality of initial solution information as the objective solution information.
Referring to fig. 2, fig. 2 shows a flowchart of a service processing method based on mixed integer linear programming according to an embodiment of the present disclosure, which specifically includes the following steps.
Step 202: and receiving a service prediction request aiming at a target service, wherein the service prediction request comprises service expected information and service constraint information corresponding to the target service.
The target service specifically refers to a specific service application scene in actual application. For example, in energy power scenarios, a combination of units problem, a transformer combination problem; dock allocation problems for ship berthing; a routing line allocation problem of a network allocation scene; channel allocation problems for satellite communications scenarios; cargo distribution problems for warehouse scenarios, and the like. It should be noted that the target service provided in the embodiment of the present specification is a mixed integer linear programming task, and preferably, the target service is a mixed 0-1 linear programming task.
The service prediction request specifically refers to a request for predicting a target service, and in practical application, the service prediction request includes service expected information and service constraint information corresponding to the target service, wherein the service expected information specifically refers to a mathematical model generated after abstraction of the target service, and the service constraint information specifically refers to constraint conditions for the service expected information. In the embodiment provided in the present specification, the service constraint information specifically includes first service constraint information and second service constraint information, where the first service constraint information includes an integer decision parameter, and the integer decision parameter specifically refers to that at least one parameter in the service expectation information is an integer.
In a specific embodiment provided in the present specification, taking a problem of allocation of a satellite communication channel in a satellite communication scenario as an example, the problem may be abstracted into a corresponding mixed integer linear programming task model, specifically, see the following target service source linear programming model 1:
minc T x+d T y
s.t.A 1 x+A 2 y=b
C 1 x+C 2 y≤d
x i ∈{L i ,U i }，i= 1，……p
y j ∈{0，1}，j= 1，……q
wherein min isc T x+d T yFor traffic intended information, i.e. objective functions, in particular in the allocation problem of satellite communication channels, the objective function is made to be "by taking the values x and y"c T x+d T y"minimum, x and y are traffic parameters,c T andd T and 4 pieces of business constraint information are included in the target business original linear programming model 1, wherein the first three pieces of business constraint information are second business constraint information, and the fourth piece of business constraint information is first business constraint information. Wherein the method comprises the steps ofA 1 、A 2 、b、C 1 、 C 2 、d、L i 、U i 、p、qThe isoparameter is specifically a constant parameter in a mixed integer linear programming task model. In the fourth business constraint information, the guarantee is neededy j Take the value of 0 or 1.
Step 204: and decomposing the business expected information into a plurality of business solution sub-information based on the business constraint information.
The embodiments provided in the present disclosure are applied to a mixed integer linear programming problem, where the mixed integer linear programming problem includes integer decision parameters, that is, at least one parameter must be an integer in the service expectation information.
Based on the above, in order to improve the solution efficiency of the service expectation information, a plurality of parameters can be generated for the service expectation information according to the service constraint information, the parameters are assigned to the service expectation information, a plurality of service solution sub-information is generated, each service solution sub-information is solved, and the optimal solution is selected from the solution result of each service solution sub-information.
Specifically, as described in the above steps, the service constraint information includes first service constraint information and second service constraint information, where the first service constraint information includes integer decision parameters, and correspondingly, the service expected information is decomposed into a plurality of service solution sub-information based on the service constraint information, including S2042-S2046:
s2042: and generating relaxation constraint information based on the first business constraint information.
When the problem of the mixed integer linear programming task model is processed, a linear programming relaxation mode is generally adopted to carry out linear relaxation on the first business constraint information, a linear programming relaxation model corresponding to the target business is obtained, and the mixed integer linear programming problem is adjusted to be a linear programming problem.
For example, as described abovey j E {0,1} is exemplified by a linear programming relaxation, denoted yj E [0,1 ] ]. The target business linear programming model 1 can be written as the following target business linear programming relaxation model 2:
minc T x+d T y
s.t.A 1 x+A 2 y=b
C 1 x+C 2 y≤d
x i ∈{L i ,U i }，i= 1，……p
y j ∈[0，1]，j= 1，……q
and the first business constraint information is subjected to linear relaxation to obtain relaxation constraint information, so that quick solution can be realized in a subsequent solving task.
S2044: and determining at least one target sampling value according to the relaxation constraint information and the second business constraint information.
After the relaxation constraint information is determined, at least one target sampling value is determined according to the relaxation constraint information and the second service constraint information, wherein the target sampling value specifically refers to a group of sampling values conforming to the first service constraint information and the second service constraint information, for example, 3 parameters exist in an objective function, then the value of 3 parameters exists in one target sampling value, and each target sampling value is obtained on the basis of the relaxation constraint information and the second constraint information and further satisfies the value of the first service constraint information.
Specifically, determining at least one target sampling value according to the relaxation constraint information and the second service constraint information includes S20442 to S20446:
s20442, determining a sampling feasible region according to the relaxation constraint information and the second business constraint information.
The sampling feasible region specifically refers to a region which is determined based on the relaxation constraint information and the second business constraint information and can be sampled, and the relaxation constraint information and the second business constraint information are both linear constraint information and are continuous, so that the corresponding sampling feasible region can be determined based on the relaxation constraint information and the second business constraint information.
Referring to fig. 3, fig. 3 shows a schematic diagram of a sampling feasible according to an embodiment of the present disclosure. As shown in fig. 3, there are a total of 3 service constraints, and the 3 service constraints form a closed region in the coordinate system, and the closed region is the sampling feasible region. The points taken in the sampling feasible domain are all points meeting the relaxation constraint information and the second business constraint information.
And S20444, determining a sampling value to be processed in the sampling feasible domain.
The to-be-processed sampling value specifically refers to a point selected in a sampling feasible domain and meeting the relaxation constraint information and the second service constraint information, and in the embodiment provided in the specification, the to-be-processed sampling value may be called an inner point.
Taking the target service linear programming relaxation model 2 as an example, in order to solve the solution of the target service linear programming relaxation model 2, it may be deformed:
The target traffic linear programming relaxation model 2 can be rewritten as:
at this time, the problem is converted into a solution of a linear equation setA’z=bIs a general solution to (a). If the linear equation set has no solution, the target business original linear programming model 1 is indicated to have no solution; if the linear equation system has a solution, a special solution can be obtainedz* A matrixBSo that an arbitrary vectorvAll satisfyz=z*+BvIs thatA’z=bIs a solution to (a). Will bez=z*+BvInstead of returning to the target traffic linear programming relaxation model 2, the target traffic linear programming relaxation model 2 may be converted into the following linear programming model 3:
min (c’) T Bv+(c’) T z*
s.t.C’Bv≤d-C’z*
L-z*≤Bv≤U-z*
solving the linear programming model 3 from the solution obtainedv* An interior point (sample to be processed) can be constructed:z 0 =z*+Bv*。
a specific solution for solving the linear programming model 3 is provided below, and in practical applications, there are other solution, which is not limited in this specification.
Based on the linear programming model 3, the following linear programming model 4 is constructed:
solving the linear programming model 4 to obtain a solution of the linear programming model 4v* Based onv* A to-be-processed can be determinedSample value (inner point):z 0 =z*+Bv*。
s20446, determining at least one target sampling value based on the sampling value to be processed.
After a sample value to be processed is determined, at least one target sample value can be determined based on the sample value to be processed, so that the selected target sample value is determined based on a sampling feasible domain, a large number of invalid sample values can be avoided, and the calculation workload is reduced.
Specifically, determining at least one target sampling value based on the sampling value to be processed includes:
determining at least one sampling direction based on the sample value to be processed;
determining at least one initial sample value in each sampling direction;
and determining at least one target sampling value according to the integer decision parameter and at least one initial sampling value.
After determining the sample value to be processed, a sampling direction may be randomly determined in a sampling feasible domain based on the sample value to be processed, and at least one initial sample value may be determined in the interview direction, see fig. 4, where fig. 4 illustrates a schematic diagram for determining an initial sample value according to an embodiment of the present disclosure. As shown in FIG. 4, a sample value to be processed is determined in the sampling feasible domainz 0 And is based onz 0 Randomly determining a sampling directionu 0 It should be noted that the sampling direction is different from positive to negative, and can be determined based on the sampling directionu 0 Randomly selecting an initial sampling value in the directionz 1 . In practical applications, a plurality of sampling directions may be determined, and a plurality of initial sampling values are collected in each sampling direction, and it should be noted that the determined initial sampling values are all determined in the sampling feasible domain determined in the above steps.
After a plurality of initial sample values are determined, each initial sample value can be determined as a target sample value. In the above steps, the first service constraint information is adjusted to the relaxation constraint information, so that the value of the initial sampling value may not satisfy the integer decision parameter in the first service constraint information. Based on this, the initial sampling value needs to be adjusted to a corresponding target sampling value according to the integer decision parameter in the first service constraint information, for example, the integer decision parameter in the initial sampling value is adjusted to a corresponding integer.
Specifically, determining at least one target sampling value according to the integer decision parameter and at least one initial sampling value includes:
determining an initial sampling value to be processed;
determining a target sampling value to be processed based on the integer decision parameter and the initial sampling value to be processed;
and under the condition that the target sampling value to be processed meets the first business constraint information and the second business constraint information, determining the target sampling value to be processed as a target sampling value.
The to-be-processed initial sampling value specifically refers to an initial sampling value which needs to be subjected to rounding operation, and the to-be-processed target sampling value specifically refers to a value determined after the to-be-processed initial sampling value is subjected to rounding processing.
In practical application, each initial sampling value needs to be adjusted to an integer according to a certain rule, and for example, a method of rounding nearby is taken as an example, see the following formula 1:
Referring to the above-mentioned formula 1,
representing a rounding down, a +.>
Representing a rounding up, i.e. when +>
The difference from its rounding down is less than +.>
The difference from this is that +.>
Rounding downwards; in addition to this, will->
And rounding upwards so as to obtain a target sampling value to be processed, which corresponds to the initial sampling value to be processed.
In practical application, after the initial sampling value to be processed is rounded, the corresponding target sampling value to be processed can be determined, and the obtained target sampling value to be processed possibly exceeds the constraint conditions of the first business constraint information and the second business constraint information, so that the target sampling value to be processed is also required to be compared with the first business constraint information and the second business constraint information, and the target sampling value to be processed can be determined to be the target sampling value under the condition that the target sampling value to be processed meets the first business constraint information and the second business constraint information.
S2046: and generating a plurality of service solving sub-information according to the at least one target sampling value and the service expected information.
After the target sampling values are determined, each target sampling value is returned to the original linear programming model of the target service, and the target sampling values can be brought into the service expected information and a plurality of service solving sub-information because the target sampling values meet the requirements of the first service constraint information and the second service constraint information.
Specifically, generating a plurality of service solution sub-information according to the at least one target sampling value and the service expected information includes:
determining a target sampling value to be processed;
and determining target service solving sub-information corresponding to the target sampling value to be processed based on the service expected information and the target sampling value to be processed.
The target sampling value to be processed specifically refers to a target sampling value to be calculated, for example, 10 target sampling values in total, and when a first target sampling value needs to be processed, the first target sampling value is the target sampling value to be processed; when the fifth target sampling value is processed, the fifth target sampling value is the target sampling value to be processed.
The target sampling value to be processed at this time can be brought into the service expected information, and target service solving sub-information corresponding to the target sampling value to be processed is generated. Further, the objective function corresponding to the objective function is included in the expected business information, for example, the objective function is included in the expected business information' min c T x+d T y"in" the "c T x+d T y"is an objective function, wherein"yAnd the 'is the parameter of the service expected information'. In the way of'yFor example, take =1 "to bring the target sample value to be processed into'y"in which target service solution sub-information min can be obtainedc T x+d T I.e. solvingc T x+d T Is a value of (2). If in the form of'yBy way of example, take the target sample value to be processed into "y"in which target service solution sub-information min can be obtainedc T x+3d T I.e. solvingc T x+3d T Is a value of (2).
And respectively bringing each target sampling value to be processed into the service expected information, and constructing and generating target service solving sub-information corresponding to each target sampling value to be processed, so that a plurality of service solving sub-information can be generated.
In another specific embodiment provided in the present specification, decomposing the service expected information into a plurality of service solution sub-information based on the service constraint information includes:
receiving the decomposition configuration information of a user, wherein the decomposition configuration information comprises the number of service solution sub-information;
and decomposing the service expected information into a plurality of service solving sub-information based on the service solving sub-information quantity and the service constraint information.
Specifically, in the process of generating a plurality of service solving sub-information, the user may control the number of service solving sub-information, and further, the user may send the decomposition configuration information to the terminal, where the decomposition configuration information includes the number of service solving sub-information, for example, the number of service solving sub-information is 1000.
When the service solving sub-information is generated, the number of the target sampling values can be determined according to the number of the service solving sub-information carried in the decomposition configuration information, so that the number of the service solving sub-information is determined.
Step 206: and respectively solving the plurality of service solving sub-information to obtain a target prediction sampling value corresponding to the target service.
After the plurality of service solving sub-information is obtained, each service solving sub-information is mutually independent, so that each service solving sub-information can be solved respectively, and the service solving sub-information is mutually independent and is not affected in the solving process. Solving each service solving sub-information to obtain the solving information corresponding to each service solving sub-information, selecting target solving information according with service expected information according to a plurality of solving information, and taking a target sampling value corresponding to the target solving information as a final target prediction sampling value.
For example, taking a total of 3 pieces of service solution sub-information as an example, the solution information 1 of the service solution sub-information 1 is 4, the solution information 2 of the service solution sub-information 2 is 7, the solution information 3 of the service solution sub-information 3 is 15, and the service expected information is a selected minimum value, the solution information 1 can be determined to be the target solution information, and meanwhile, the target sampling value 1 corresponding to the solution information 1 can be determined to be the target prediction sampling value.
Specifically, solving the plurality of service solving sub-information respectively to obtain a target prediction sampling value corresponding to the target service, including S2062-S2064:
s2062, calculating the sub-information of each service solution to obtain the initial solution information corresponding to each target sampling value.
In practical application, the data volume of the target sampling values may be large, so that the number of the generated service solving sub-information is also large, and because the calculation among the plurality of the target sampling values is independent and does not affect each other, the plurality of the service solving sub-information is independent and does not affect each other in the calculation solving process, and the overall calculation speed can be improved through the parallel calculation processing among the plurality of the service solving sub-information.
Specifically, calculating the solution sub-information of each service to obtain the initial solution information corresponding to each target sampling value, including:
generating a sub-model to be processed based on the target service solving sub-information;
and solving the sub-model to be processed to obtain initial solving information corresponding to the target sampling value to be processed.
In order to facilitate the independent calculation of each service solving sub-information, a corresponding sub-model to be processed can be generated according to the service solving sub-information, wherein the sub-model to be processed refers to a mathematical processing model corresponding to each service solving sub-information and related to a target service, and the model result obtained by solving the sub-model to be processed can be used as initial solving information corresponding to a target sampling value to be processed.
In practical application, because each sub-model to be processed is processed independently and in parallel, in order to improve the processing efficiency of the model, the calculation time can be shortened by the idea of distributed processing.
Specifically, solving the sub-model to be processed to obtain initial solution information corresponding to a target sampling value to be processed, including:
determining a target computing child node;
and sending the sub-model to be processed to the target computing sub-node, and receiving initial solving information returned by the target computing sub-node.
In a specific embodiment provided in the present specification, each sub-model to be processed may be further distributed to computing nodes for parallel processing, where a target computing node specifically refers to a node having enough computing power resources to calculate a result of the sub-model to be processed, where the target computing node includes, but is not limited to, a computer, a virtual machine, a CPU core, a process, a thread, and the like, and the computing node has a capability of receiving model data of the sub-model to be processed, executing a linear programming solver to obtain a solution of the sub-model, and sending initial solution information of the sub-model to be processed to the execution body through a channel such as a storage channel, a network, and the like.
The sub-model to be processed is distributed to a plurality of target computing sub-nodes for processing, so that computing resources can be fully utilized, and the waiting total time of model computing is reduced.
S2064, determining a target prediction sampling value corresponding to the target service based on each piece of initial solving information.
After each initial solution information is determined, the final target prediction sample value of the target service can be further determined. Specifically, determining the target prediction sampling value corresponding to the target service based on each piece of initial solution information includes:
determining an expected value of the service expected information;
determining target solving information in each piece of initial solving information based on the expected value;
and determining a target sampling value corresponding to the target solving information as a target prediction sampling value.
In practical application, the business expectation information includes the corresponding objective function, for example, in the business expectation information "minc T x+d T y"in" the "c T x+d T yThe "min" is the expected value of the expected information of the business, namely the expected value of the expected target function is minimum, and the standard for determining the target prediction sampling value can be obtained through the expected value of the expected information of the business. For example, when the expected value is "take small value", the minimum value in the initial solution information is taken as the target solution information, and the target sampling value corresponding to the target solution information is taken as the target prediction sampling value of the target service.
In a specific embodiment provided in the present specification, the service expectation information "minc T x+d T y"for example, in the case of target sample value"yCase of =1'The lower part of the upper part is provided with a lower part,c T x+d T is m, at a target sampling value'yIn the case of =3″,c T x+3d T the initial solution information m is the target solution information, and the target sampling value corresponding to the target solution information m is the target solution information after comparison "y=1 ", i.e. the target predicted sample value.
In a specific embodiment provided in the present specification, the method further includes:
and in the process of solving the sub-model to be processed, sending the solving information of the sub-model to be processed to the front end, wherein the solving information comprises at least one of solving efficiency, solving progress and solving residual time.
In practical application, the calculation between the sub-models to be processed is independent and does not affect each other, and in the solving process, each sub-model to be processed can send the solving information generated in the solving process to the front end of the user for display, so that the user can know the processing condition between a plurality of sub-models to be processed, and the solving information includes but is not limited to solving efficiency, solving progress, solving residual time and the like.
In another specific embodiment provided in the present specification, there may be a case where there are at least two generated target prediction sampling values, based on which the method provided in the present specification further includes:
transmitting the at least two target prediction sampling values to the front end under the condition that the target prediction sampling values are at least two;
and receiving a selection instruction sent by a user at the front end, and determining a final target prediction sampling value according to the selection instruction.
In the mixed integer linear programming problem corresponding to the target service, an optimal solution needs to be found, and usually, only one group of optimal solutions exists. There may be cases where there are at least two optimal solutions, i.e., there are at least two target predicted sample values. At this time, at least two target prediction sampling values may be sent to the front end for selection by the user, and the user may select one of the at least two target prediction sampling values as a final target prediction sampling value.
In another embodiment provided in the present specification, one of the at least two target prediction sampling values may be optionally selected as the final target prediction sampling value.
Further, the service prediction request for the target service may be multiple times, and accordingly, the method provided in the present specification further includes:
And taking the target prediction sampling value as a candidate prediction value of the target service, wherein the candidate prediction value is used for providing selection for a user when the service prediction request is received again.
In practical applications, it may be necessary to periodically predict the target service, for example, in the group prediction service, predicting the group combination for the next day every day; in the scenario of ship berthing prediction traffic, predictions are provided for dock allocation every preset time period.
In this case, the current target prediction sampling value may be used as a candidate prediction value of a subsequent service prediction task, and the user may continue to use the target prediction sampling value of the current prediction task in the next service prediction task, or may use the target prediction sampling value of the current prediction task as a candidate prediction value of the next service prediction task.
According to the method and the device, the mixed integer linear programming problem is decomposed into the plurality of business solving sub-information according to the business constraint information, the plurality of business solving sub-information is processed in parallel, the solving result corresponding to each business solving sub-information is obtained, the business solving sub-information is independent, the parallel calculation can accelerate the solving efficiency of the business solving sub-information, so that the solving speed of the mixed integer linear programming problem is improved, finally, the target prediction sampling value of the mixed integer linear programming problem is obtained according to the result of each business solving sub-information, the optimal solution can be rapidly screened out in the limited solving result, the solving speed is further accelerated, and the calculating time is saved.
Furthermore, in the processing process, the mixed integer linear programming problem is firstly converted into the linear programming problem, and the target sampling value is determined in the relaxation constraint information and the second business constraint information, so that the effectiveness of the target sampling value is ensured, the resource waste caused by invalid sampling points is greatly reduced, the subsequent calculation efficiency is improved, each target sampling point is mutually independent, a plurality of business solving sub-information is generated according to the target sampling value, and the plurality of business solving sub-information supports large-scale parallel calculation, so that the solving speed can be increased, and the calculation time is saved.
The following describes, with reference to fig. 5, an example of application of the service processing method based on mixed integer linear programming provided in the present specification to dock berth scene allocation. Fig. 5 shows a process flow chart of a business processing method based on mixed integer linear programming according to an embodiment of the present disclosure, which specifically includes the following steps.
Step 502: a berth prediction request for allocation of berths for a vessel is received, wherein the berth prediction request is a mixed integer linear programming problem.
Step 504: analyzing the berth prediction request to obtain berth expected information, first constraint information and second constraint information carried in the request, wherein the first constraint information requires the number of berths to be an integer.
Step 506: and linearly relaxing the first constraint information to generate relaxation constraint information.
Step 508: and determining a sampling feasible domain according to the relaxation constraint information and the second constraint information, and determining a sampling value to be processed in the sampling feasible domain.
Step 510: at least one sampling direction is determined based on the sample values to be processed, and at least one initial sample value is determined in each sampling direction.
Step 512: at least one target sample value is determined based on the requirement that the number of berths be an integer and the at least one initial sample value.
Step 514: and generating a sub-model to be processed corresponding to each target sampling value according to the berth expected information and each target sampling value.
Step 516: and determining a target computing node corresponding to each sub-model to be processed, and sending each sub-model to be processed to the corresponding target computing node.
Step 518: and receiving initial solving information returned by each target computing node.
Step 520: and selecting a minimum value from each piece of initial solution information as target solution information.
Step 522: and determining a target sampling value corresponding to the target solving information as predicted berth allocation information corresponding to the berth prediction request.
According to the method provided by the application, the berth allocation problem of the mixed integer linear programming is firstly converted into the linear programming problem, the target sampling value is determined in the relaxation constraint information and the second business constraint information, the effectiveness of the target sampling value is guaranteed, the resource waste caused by invalid sampling points is greatly reduced, the subsequent calculation efficiency is improved, each target sampling point is mutually independent, and the large-scale parallel calculation is supported, so that the solving speed can be increased, and the calculation time is saved.
Referring to fig. 6, fig. 6 shows a flowchart of a unit combination prediction method based on mixed integer linear programming according to an embodiment of the present disclosure, which specifically includes the following steps:
step 602: and receiving a service prediction request for a unit combination prediction service, wherein the service prediction request comprises service expected information and service constraint information corresponding to the unit combination prediction service.
Step 604: and decomposing the business expected information into a plurality of business solution sub-information based on the business constraint information.
Step 606: and respectively solving the plurality of service solving sub-information to obtain the prediction unit combination corresponding to the unit combination prediction service.
In a specific embodiment provided in the present specification, the service constraint information includes first service constraint information and second service constraint information, and the first service constraint information includes an integer decision parameter;
decomposing the business expectation information into a plurality of business solution sub-information based on the business constraint information, including:
generating relaxation constraint information based on the first business constraint information;
determining at least one target sampling unit combination according to the relaxation constraint information and the second business constraint information;
and generating a plurality of service solution sub-information according to the at least one target sampling unit combination and the service expected information.
In a specific embodiment provided in the present specification, solving a plurality of service solution sub-information to obtain a prediction unit combination corresponding to the unit combination prediction service includes:
calculating the sub-information of each service solution to obtain initial solution information corresponding to each target sampling unit combination;
and determining a predicted unit combination corresponding to the unit combination prediction service based on each initial solution information.
In a specific embodiment provided in the present specification, determining at least one target sampling set combination according to the relaxation constraint information and the second service constraint information includes:
determining a sampling feasible region according to the relaxation constraint information and the second business constraint information;
determining a sampling value to be processed in the sampling feasible domain;
at least one target sample set combination is determined based on the sample values to be processed.
In a specific embodiment provided in the present specification, determining at least one target sampling group combination based on the sample values to be processed includes:
determining at least one sampling direction based on the sample value to be processed;
determining at least one initial sample value in each sampling direction;
and determining at least one target sampling unit combination according to the integer decision parameter and at least one initial sampling value.
In a specific embodiment provided in the present specification, determining at least one target sampling set combination according to the integer decision parameter and at least one initial sampling value includes:
determining an initial sampling value to be processed;
determining a target sampling unit combination to be processed based on the integer decision parameter and the initial sampling value to be processed;
And determining the target sampling unit combination to be processed as a target sampling unit combination under the condition that the target sampling unit combination to be processed meets the first business constraint information and the second business constraint information.
In a specific embodiment provided in the present specification, generating a plurality of service solution sub-information according to the at least one target sampling group combination and the service expectation information includes:
determining a target sampling unit combination to be processed;
and determining unit combination prediction service solving sub-information corresponding to the target sampling unit combination to be processed based on the service expected information and the target sampling unit combination to be processed.
Calculating the sub-information of each service solution to obtain the initial solution information corresponding to each target sampling unit combination, including:
generating a sub-model to be processed based on the unit combination prediction service solution sub-information;
and solving the sub-model to be processed to obtain initial solving information corresponding to the combination of the target sampling units to be processed.
In a specific embodiment provided in the present specification, the method further includes:
and in the process of solving the sub-model to be processed, sending the solving information of the sub-model to be processed to the front end, wherein the solving information comprises at least one of solving efficiency, solving progress and solving residual time.
In a specific embodiment provided in the present specification, decomposing the service expected information into a plurality of service solution sub-information based on the service constraint information includes:
receiving the decomposition configuration information of a user, wherein the decomposition configuration information comprises the number of service solution sub-information;
and decomposing the service expected information into a plurality of service solving sub-information based on the service solving sub-information quantity and the service constraint information.
In a specific embodiment provided in the present specification, the method further includes:
transmitting the at least two prediction unit combinations to a front end under the condition that the number of the prediction unit combinations is at least two;
and receiving a selection instruction sent by a user at the front end, and determining a final prediction unit combination according to the selection instruction.
In a specific embodiment provided in the present specification, the method further includes:
and taking the prediction unit combination as a candidate prediction value of the unit combination prediction service, wherein the candidate prediction value is used for providing a selection for a user when a service prediction request is received again.
According to the method provided by the application, the unit combination prediction problem of the mixed integer linear programming is firstly converted into the linear programming problem, the target sampling unit combination is determined in the relaxation constraint information and the second business constraint information, the effectiveness of the target sampling unit combination is guaranteed, the resource waste caused by invalid sampling points is greatly reduced, the subsequent calculation efficiency is improved, each target sampling point is mutually independent, and the large-scale parallel calculation is supported, so that the solving speed can be increased, and the calculation time is saved.
Corresponding to the above embodiment of the service processing method based on mixed integer linear programming, the present disclosure further provides an embodiment of a service processing device based on mixed integer linear programming, and fig. 7 shows a schematic structural diagram of a service processing device based on mixed integer linear programming according to one embodiment of the present disclosure. As shown in fig. 7, the apparatus includes:
a receiving module 702, configured to receive a service prediction request for a target service, where the service prediction request includes service expected information and service constraint information corresponding to the target service;
a generation module 704 configured to decompose the business desired information into a plurality of business solution sub-information based on the business constraint information;
and the prediction module 706 is configured to solve the plurality of service solution sub-information respectively to obtain a target prediction sampling value corresponding to the target service.
Optionally, the service constraint information includes first service constraint information and second service constraint information, and the first service constraint information includes integer decision parameters;
the generating module 704 is further configured to:
generating relaxation constraint information based on the first business constraint information;
Determining at least one target sampling value according to the relaxation constraint information and the second business constraint information;
and generating a plurality of service solving sub-information according to the at least one target sampling value and the service expected information.
Optionally, the prediction module 706 is further configured to:
calculating the sub-information of each service solution to obtain the initial solution information corresponding to each target sampling value;
and determining a target prediction sampling value corresponding to the target service based on each piece of initial solving information.
Optionally, the generating module 704 is further configured to:
determining a sampling feasible region according to the relaxation constraint information and the second business constraint information;
determining a sampling value to be processed in the sampling feasible domain;
at least one target sample value is determined based on the sample value to be processed.
Optionally, the generating module 704 is further configured to:
determining at least one sampling direction based on the sample value to be processed;
determining at least one initial sample value in each sampling direction;
and determining at least one target sampling value according to the integer decision parameter and at least one initial sampling value.
Optionally, the generating module 704 is further configured to:
Determining an initial sampling value to be processed;
determining a target sampling value to be processed based on the integer decision parameter and the initial sampling value to be processed;
and under the condition that the target sampling value to be processed meets the first business constraint information and the second business constraint information, determining the target sampling value to be processed as a target sampling value.
The generating module 704 is further configured to:
determining a target sampling value to be processed;
and determining target service solving sub-information corresponding to the target sampling value to be processed based on the service expected information and the target sampling value to be processed.
Optionally, the prediction module 706 is further configured to:
generating a sub-model to be processed based on the target service solving sub-information;
and solving the sub-model to be processed to obtain initial solving information corresponding to the target sampling value to be processed.
Optionally, the apparatus further includes:
and the feedback module is configured to send the solving information of the sub-model to be processed to the front end in the process of solving the sub-model to be processed, wherein the solving information comprises at least one of solving efficiency, solving progress and solving residual time.
Optionally, the generating module 704 is further configured to:
receiving the decomposition configuration information of a user, wherein the decomposition configuration information comprises the number of service solution sub-information;
and decomposing the service expected information into a plurality of service solving sub-information based on the service solving sub-information quantity and the service constraint information.
Optionally, the apparatus further includes:
the information sending module is configured to send the at least two target prediction sampling values to the front end when the target prediction sampling values are at least two;
and the selection module is configured to receive a selection instruction sent by a user at the front end and determine a final target prediction sampling value according to the selection instruction.
Optionally, the apparatus further includes:
and the candidate module is configured to take the target prediction sampling value as a candidate prediction value of the target service, wherein the candidate prediction value is used for providing selection for a user when the service prediction request is received again.
The service processing device based on mixed integer linear programming provided by the embodiment of the specification comprises a service prediction request for a target service, wherein the service prediction request comprises service expected information corresponding to the target service, first service constraint information and second service constraint information, and the first service constraint information comprises integer decision parameters; generating relaxation constraint information based on the first business constraint information; determining at least one target sampling value according to the relaxation constraint information and the second business constraint information; and calculating an initial service predicted value corresponding to each target sampling value according to the service expected information, and determining the target predicted sampling value corresponding to the target service based on each initial service predicted value. According to the method provided by the application, the mixed integer linear programming problem is firstly converted into the linear programming problem, the target sampling value is determined in the relaxation constraint information and the second business constraint information, the effectiveness of the target sampling value is guaranteed, the resource waste caused by invalid sampling points is greatly reduced, the subsequent calculation efficiency is improved, each target sampling point is mutually independent, and the large-scale parallel calculation is supported, so that the solving speed can be increased, and the calculation time is saved.
The above is an exemplary scheme of a service processing apparatus based on mixed integer linear programming of the present embodiment. It should be noted that, the technical solution of the service processing apparatus based on mixed integer linear programming and the technical solution of the service processing method based on mixed integer linear programming belong to the same concept, and details of the technical solution of the service processing apparatus based on mixed integer linear programming, which are not described in detail, can be referred to the description of the technical solution of the service processing method based on mixed integer linear programming.
Referring to fig. 8, fig. 8 shows a schematic diagram of an apparatus for a unit combination based on mixed integer linear programming according to an embodiment of the present disclosure. As shown in fig. 8, the prediction problem of the crew combination is abstracted into crew combination model data, and the crew combination model data includes service expectation information, first service constraint information and second service constraint information of the crew combination prediction service, wherein the first service constraint information includes integer decision parameters.
The set combination model data is input to the interior point calculator 802, the interior point calculator 802 performs linear relaxation on the first service constraint information in the set combination model to obtain relaxation constraint information, and generates a sampling feasible domain according to the relaxation constraint information and the second service constraint information, a feasible solution of the linear programming relaxation model is determined in the sampling feasible domain, and an interior point (sampling value to be processed) is constructed based on the feasible solution.
After the interior points are built, the interior points are input into the random sampler 804, at least one sampling direction is determined based on the interior points in the random sampler 804, a line segment with two ends touching the boundary of the sampling feasible region can be constructed in the forward and backward directions of the sampling direction, at least one initial sampling value is determined on the line segment, the steps of determining the sampling direction and determining the initial sampling values are repeated until a preset number of initial sampling values are obtained.
Each initial sampling value is input into the sub-problem constructor 806, the sub-problem constructor 806 takes the integer decision parameter to the integer value according to a preset rule, obtains a target sampling value to be processed corresponding to each initial sampling value, and reserves the target sampling value to be processed meeting the first service constraint information and the second service constraint information as a final target sampling value. And the sub-problem constructor 806 returns the set combined model according to each target sampling value to generate a sub-model to be processed corresponding to each target sampling value. The solution problem of the crew combination model is divided into a plurality of optimization solution sub-problems by the sub-problem constructor 806.
After each sub-model to be processed is created, the sub-problem distributor transmits each sub-model to be processed to a computing node provided with a linear programming solver through a storage channel, a network channel and the like.
The compute nodes are any resource with sufficient computing power, including but not limited to computers, virtual machines, CPU cores, processes, threads, etc., that are capable of receiving the sub-models to be processed, executing a linear programming solver to solve the solutions of the sub-models to be processed, and transmitting the solutions of each sub-model to be processed to the summary verifier 810 via storage, network, etc.
The summarizing verifier 810 is responsible for receiving solutions generated by a plurality of computing nodes, recording and updating the optimal solutions which are currently collected, judging whether a preset termination condition is met, and if not, continuing to execute the random sampler 804 and subsequent links; if so, outputting the solution recorded currently as a feasible solution of the unit combination model.
It should be noted that the preset termination condition refers to a termination condition set manually, for example, the preset termination condition may be set to 3 rounds of calculation; and may be configured to solve for a feasible solution, and so on. In the embodiment provided in the present specification, the specific content of the preset termination condition is not limited.
Referring to fig. 9, fig. 9 illustrates a service processing system based on mixed integer linear programming according to an embodiment of the present disclosure, where the system may include an end-side device 901 and a cloud-side device 902, where the end-side device 901 is configured to send a service prediction request to the Yun Ceshe device 902 and receive a target prediction sampling value sent by the cloud-side device 902; the cloud testing device 902 is configured to perform an operation of solving a target prediction sampling value for a received service prediction request for a target service.
The terminal side device 901 is configured to generate a service prediction request for a target service, and send the service prediction request to the cloud measurement device, where the service prediction request includes service expected information and service constraint information corresponding to the target service;
the cloud testing device 902 is configured to receive the service prediction request, decompose the service expected information into a plurality of service solution sub-information based on the service constraint information, respectively solve the plurality of service solution sub-information, obtain a target prediction sampling value corresponding to the target service, and send the target prediction sampling value to the end-side device.
Cloud-side device 902 may be a central cloud device of a distributed cloud architecture, end-side device 901 may be an edge cloud device of the distributed cloud architecture, cloud-side device 902 and end-side device 901 may be service end devices such as a conventional server, a cloud server or a server array, or may be terminal devices, which is not limited in this embodiment of the present disclosure. Moreover, cloud-side device 902 provides ultra-strong computing and storage capabilities, remote from the user; while the end side device 901 is deployed in a large range, closer to the user. The end side device 901 is an expansion of the cloud side device 902, and can sink the computing capability of the cloud side device 902 to the end side device 901, and solve the service requirements which cannot be met in the centralized cloud computing mode through the integration and collaborative management of the end cloud.
Fig. 10 illustrates a block diagram of a computing device 1000 provided in accordance with one embodiment of the present description. The components of the computing device 1000 include, but are not limited to, a memory 1010 and a processor 1020. Processor 1020 is coupled to memory 1010 via bus 1030 and database 1050 is used to store data.
Computing device 1000 also includes access device 1040, which access device 1040 enables computing device 1000 to communicate via one or more networks 1060. Examples of such networks include public switched telephone networks (PSTN, public SwitchedTelephone Network), local area networks (LAN, local AreaNetwork), wide Area Networks (WAN), personal area networks (PAN, personal AreaNetwork), or combinations of communication networks such as the internet. The access device 1040 may include one or more of any type of network interface, wired or wireless, such as a network interface card (NIC, networkinterface controller), such as an IEEE802.11 wireless local area network (WLAN, wireless LocalArea Network) wireless interface, a worldwide interoperability for microwave access (Wi-MAX, worldwideInteroperability for Microwave Access) interface, an ethernet interface, a universal serial bus (USB, universal Serial Bus) interface, a cellular network interface, a bluetooth interface, near field communication (NFC, near FieldCommunication).
In one embodiment of the present description, the above-described components of computing device 1000, as well as other components not shown in FIG. 10, may also be connected to each other, such as by a bus. It should be understood that the block diagram of the computing device illustrated in FIG. 10 is for exemplary purposes only and is not intended to limit the scope of the present description. Those skilled in the art may add or replace other components as desired.
Computing device 1000 may be any type of stationary or mobile computing device including a mobile computer or mobile computing device (e.g., tablet, personal digital assistant, laptop, notebook, netbook, etc.), mobile phone (e.g., smart phone), wearable computing device (e.g., smart watch, smart glasses, etc.), or other type of mobile device, or a stationary computing device such as a desktop computer or personal computer (PC, personalComputer). Computing device 1000 may also be a mobile or stationary server.
Wherein the processor 1020 is configured to execute computer-executable instructions that, when executed by the processor, perform the steps of the data processing method described above. The foregoing is a schematic illustration of a computing device of this embodiment. It should be noted that, the technical solution of the computing device and the technical solution of the service processing method based on the mixed integer linear programming or the unit combination prediction method based on the mixed integer linear programming belong to the same concept, and details of the technical solution of the computing device, which are not described in detail, can be referred to the description of the technical solution of the service processing method based on the mixed integer linear programming or the unit combination prediction method based on the mixed integer linear programming.
An embodiment of the present disclosure further provides a computer readable storage medium storing computer executable instructions that, when executed by a processor, implement the steps of the above-described hybrid integer linear programming-based service processing method or hybrid integer linear programming-based unit combination prediction method.
The above is an exemplary version of a computer-readable storage medium of the present embodiment. It should be noted that, the technical solution of the storage medium and the technical solution of the service processing method based on the mixed integer linear programming or the unit combination prediction method based on the mixed integer linear programming belong to the same concept, and details of the technical solution of the storage medium which are not described in detail can be referred to the description of the technical solution of the service processing method based on the mixed integer linear programming or the unit combination prediction method based on the mixed integer linear programming.
An embodiment of the present disclosure further provides a computer program, where the computer program when executed in a computer causes the computer to perform the steps of the above-mentioned business processing method based on mixed integer linear programming or the unit combination prediction method based on mixed integer linear programming.
The above is an exemplary version of a computer program of the present embodiment. It should be noted that, the technical solution of the computer program and the technical solution of the service processing method based on the mixed integer linear programming or the unit combination prediction method based on the mixed integer linear programming belong to the same concept, and details of the technical solution of the computer program, which are not described in detail, can be referred to the description of the technical solution of the service processing method based on the mixed integer linear programming or the unit combination prediction method based on the mixed integer linear programming.
The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
The computer instructions include computer program code that may be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random AccessMemory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the computer readable medium contains content that can be appropriately scaled according to the requirements of jurisdictions in which such content is subject to legislation and patent practice, such as in certain jurisdictions in which such content is subject to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.
It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of combinations of actions, but it should be understood by those skilled in the art that the embodiments are not limited by the order of actions described, as some steps may be performed in other order or simultaneously according to the embodiments of the present disclosure. Further, those skilled in the art will appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily all required for the embodiments described in the specification.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.
The preferred embodiments of the present specification disclosed above are merely used to help clarify the present specification. Alternative embodiments are not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the teaching of the embodiments. The embodiments were chosen and described in order to best explain the principles of the embodiments and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention. This specification is to be limited only by the claims and the full scope and equivalents thereof.
Claims (18)
1. A business processing method based on mixed integer linear programming comprises the following steps:
receiving a service prediction request aiming at a target service, wherein the service prediction request comprises service expected information and service constraint information corresponding to the target service, the service expected information comprises a mathematical model corresponding to the target service, and the service constraint information comprises constraint conditions of the service expected information;
decomposing the business desire information into a plurality of business solution sub-information based on the business constraint information, wherein decomposing the business desire information into a plurality of business solution sub-information based on the business constraint information comprises determining at least one target sampling value based on the business constraint information, and generating a plurality of business solution sub-information according to the at least one target sampling value and the business desire information;
and respectively solving the plurality of service solving sub-information to obtain a target prediction sampling value corresponding to the target service.
2. The method of claim 1, the traffic constraint information comprising first traffic constraint information and second traffic constraint information, the first traffic constraint information comprising integer decision parameters;
Determining at least one target sample value based on the traffic constraint information, comprising:
generating relaxation constraint information based on the first business constraint information;
and determining at least one target sampling value according to the relaxation constraint information and the second business constraint information.
3. The method of claim 1, respectively solving a plurality of service solution sub-information to obtain a target prediction sampling value corresponding to the target service, comprising:
calculating the sub-information of each service solution to obtain the initial solution information corresponding to each target sampling value;
and determining a target prediction sampling value corresponding to the target service based on each piece of initial solving information.
4. The method of claim 2, determining at least one target sample value from the relaxation constraint information and the second traffic constraint information, comprising:
determining a sampling feasible region according to the relaxation constraint information and the second business constraint information;
determining a sampling value to be processed in the sampling feasible domain;
at least one target sample value is determined based on the sample value to be processed.
5. The method of claim 4, determining at least one target sample value based on the sample value to be processed, comprising:
Determining at least one sampling direction based on the sample value to be processed;
determining at least one initial sample value in each sampling direction;
and determining at least one target sampling value according to the integer decision parameter and at least one initial sampling value.
6. The method of claim 5, determining at least one target sample value from the integer decision parameter and at least one initial sample value, comprising:
determining an initial sampling value to be processed;
determining a target sampling value to be processed based on the integer decision parameter and the initial sampling value to be processed;
and under the condition that the target sampling value to be processed meets the first business constraint information and the second business constraint information, determining the target sampling value to be processed as a target sampling value.
7. The method of claim 1, generating a plurality of business solution sub-information from the at least one target sample value and the business desired information, comprising:
determining a target sampling value to be processed;
and determining target service solving sub-information corresponding to the target sampling value to be processed based on the service expected information and the target sampling value to be processed.
8. The method of claim 7, calculating each service solution sub-information to obtain initial solution information corresponding to each target sample value, comprising:
Generating a sub-model to be processed based on the target service solving sub-information;
and solving the sub-model to be processed to obtain initial solving information corresponding to the target sampling value to be processed.
9. The method of claim 8, further comprising:
and in the process of solving the sub-model to be processed, sending the solving information of the sub-model to be processed to the front end, wherein the solving information comprises at least one of solving efficiency, solving progress and solving residual time.
10. The method of claim 1, decomposing the business desire information into a plurality of business solution sub-information based on the business constraint information, comprising:
receiving the decomposition configuration information of a user, wherein the decomposition configuration information comprises the number of service solution sub-information;
and decomposing the service expected information into a plurality of service solving sub-information based on the service solving sub-information quantity and the service constraint information.
11. The method of claim 1, further comprising:
transmitting the at least two target prediction sampling values to the front end under the condition that the target prediction sampling values are at least two;
and receiving a selection instruction sent by a user at the front end, and determining a final target prediction sampling value according to the selection instruction.
12. The method of claim 11, further comprising:
and taking the target prediction sampling value as a candidate prediction value of the target service, wherein the candidate prediction value is used for providing selection for a user when the service prediction request is received again.
13. A unit combination prediction method based on mixed integer linear programming comprises the following steps:
receiving a service prediction request for a unit combination prediction service, wherein the service prediction request comprises service expected information and service constraint information corresponding to the unit combination prediction service, the service expected information comprises a mathematical model corresponding to the unit combination prediction service, and the service constraint information comprises constraint conditions of the service expected information;
decomposing the business desire information into a plurality of business solution sub-information based on the business constraint information, wherein decomposing the business desire information into a plurality of business solution sub-information based on the business constraint information comprises determining at least one target sampling value based on the business constraint information, and generating a plurality of business solution sub-information according to the at least one target sampling value and the business desire information;
and respectively solving the plurality of service solving sub-information to obtain the prediction unit combination corresponding to the unit combination prediction service.
14. The method of claim 13, the method further comprising:
and in the process of solving the business solving sub-information, sending the solving information of the business solving sub-information to the front end, wherein the solving information comprises at least one of solving efficiency, solving progress and solving residual time.
15. The method of claim 13, further comprising:
transmitting the at least two prediction unit combinations to a front end under the condition that the number of the prediction unit combinations is at least two;
and receiving a selection instruction sent by a user at the front end, and determining a final prediction unit combination according to the selection instruction.
16. A business processing system based on mixed integer linear programming, comprising:
the terminal side equipment is used for generating a service prediction request aiming at a target service and sending the service prediction request to the cloud testing equipment, wherein the service prediction request comprises service expected information and service constraint information corresponding to the target service, the service expected information comprises a mathematical model corresponding to the target service, and the service constraint information comprises constraint conditions of the service expected information;
the cloud testing device is configured to receive the service prediction request, decompose the service expected information into a plurality of service solving sub-information based on the service constraint information, respectively solve the plurality of service solving sub-information, obtain a target prediction sampling value corresponding to the target service, and send the target prediction sampling value to the end-side device, where decomposing the service expected information into the plurality of service solving sub-information based on the service constraint information includes determining at least one target sampling value based on the service constraint information, and generating a plurality of service solving sub-information according to the at least one target sampling value and the service expected information.
17. A computing device, comprising:
a memory and a processor;
the memory is configured to store computer executable instructions and the processor is configured to execute the computer executable instructions which, when executed by the processor, perform the steps of the method of any of claims 1-12 or 13-15.
18. A computer readable storage medium storing computer executable instructions which when executed by a processor perform the steps of the method of any one of claims 1-12 or 13-15.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310126641.0A CN115829169B (en) | 2023-02-10 | 2023-02-10 | Business processing method and device based on mixed integer linear programming |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310126641.0A CN115829169B (en) | 2023-02-10 | 2023-02-10 | Business processing method and device based on mixed integer linear programming |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115829169A CN115829169A (en) | 2023-03-21 |
| CN115829169B true CN115829169B (en) | 2023-05-16 |
Family
ID=85521739
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310126641.0A Active CN115829169B (en) | 2023-02-10 | 2023-02-10 | Business processing method and device based on mixed integer linear programming |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115829169B (en) |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060112049A1 (en) * | 2004-09-29 | 2006-05-25 | Sanjay Mehrotra | Generalized branching methods for mixed integer programming |
| US7873530B2 (en) * | 2008-04-28 | 2011-01-18 | International Business Machines Corporation | Method and system for solving stochastic linear programs with conditional value at risk constraints |
| US8676737B2 (en) * | 2009-10-05 | 2014-03-18 | International Business Machines Corporation | Detecting missing cases in business rules |
| US9418046B2 (en) * | 2014-05-14 | 2016-08-16 | International Business Machines Corporation | Price-and-branch algorithm for mixed integer linear programming |
| CN115220882A (en) * | 2021-04-19 | 2022-10-21 | 华为技术有限公司 | Data processing method and device |
| CN115496247A (en) * | 2021-06-18 | 2022-12-20 | 华为技术有限公司 | Service data processing method and device |
| CN115685892A (en) * | 2021-07-30 | 2023-02-03 | 华为技术有限公司 | Method, apparatus, device, storage medium and program product for industrial planning |
| CN114237835A (en) * | 2021-09-30 | 2022-03-25 | 华为技术有限公司 | Task solving method and device |
-
2023
- 2023-02-10 CN CN202310126641.0A patent/CN115829169B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115829169A (en) | 2023-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112633316B (en) | Load prediction method and device based on boundary estimation theory | |
| Mencía et al. | Genetic algorithms for the scheduling problem with arbitrary precedence relations and skilled operators | |
| CN108365969B (en) | Self-adaptive service combination method based on wireless sensor network | |
| CN113240226A (en) | Resource distribution method and device between satellite ground stations | |
| CN111768004A (en) | Model self-adaption method and system based on intelligent computing framework | |
| CN110334436B (en) | Data processing method and device | |
| CN117032938B (en) | Operator parallel scheduling method and device, electronic equipment and storage medium | |
| CN114741075A (en) | Task optimization method and device | |
| Colonetti et al. | A novel cooperative multi-search benders decomposition for solving the hydrothermal unit-commitment problem | |
| CN115829169B (en) | Business processing method and device based on mixed integer linear programming | |
| Li et al. | Robust transmission constrained unit commitment: a column merging method | |
| CN112766402A (en) | Algorithm selection method and device and electronic equipment | |
| Guo et al. | Data-driven robust optimization for contextual vehicle rebalancing in on-demand ride services under demand uncertainty | |
| CN117236427A (en) | Decision constraint rule characterization method for task actions | |
| CN115271130B (en) | Dynamic scheduling method and system for maintenance order of ship main power equipment | |
| Deng et al. | Neural Regret-Matching for Distributed Constraint Optimization Problems. | |
| CN116341634A (en) | Training method and device for neural structure search model and electronic equipment | |
| CN113806624B (en) | Data processing method and device | |
| Li et al. | A genetic algorithm enhanced automatic data flow management solution for facilitating data intensive applications in the cloud | |
| CN114707667A (en) | Data-driven automatic model training and application system | |
| CN114417022A (en) | Model training method, data processing method and device | |
| CN115859693B (en) | Data processing method and device | |
| Zhang et al. | Deploying GIS services into the edge: A study from performance evaluation and optimization viewpoint | |
| CN115827171B (en) | Cloud parameter adjusting system, parameter adjusting method and parameter adjusting system | |
| Chakraborty et al. | Multi-criterial Offloading Decision Making in Green Mobile Cloud Computing |
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 |