WO2022087962A1

WO2022087962A1 - Simulation-based closed-loop aps scheduling optimization method and system, and storage medium

Info

Publication number: WO2022087962A1
Application number: PCT/CN2020/124827
Authority: WO
Inventors: 廖梓博; 郁彦彬
Original assignee: 西门子股份公司; 西门子（中国）有限公司
Priority date: 2020-10-29
Filing date: 2020-10-29
Publication date: 2022-05-05
Also published as: CN116529741A

Abstract

A simulation-based closed-loop APS scheduling optimization method and system, and a storage medium. The method comprises: determining APS rules selected for scheduling (101); determining weight configuration set for the selected APS rules (102); generating a scheduling Gantt chart according to the weight configuration of the APS rules (103); deriving an order sequence corresponding to the scheduling Gantt chart (104); loading the order sequence and obtained preset simulation model configuration data into a basic simulation model of a factory design simulation software to obtain a scheduling simulation model (105); running the scheduling simulation model and evaluating to acquire key performance indicator data (106); exporting the key performance indicator data as simulation results (107); determining whether the simulation results meet needs (108); if yes, outputting the corresponding scheduling Gantt chart (109); otherwise, returning to the step of selecting the APS rules for scheduling. The present method can provide a more feasible scheduling solution for factory production.

Description

Simulation-based closed-loop APS scheduling optimization method, system and storage medium

technical field

The present application relates to the field of digitization, in particular to a simulation-based closed-loop advanced planning and scheduling software (APS) scheduling optimization method, system and storage medium.

Background technique

Currently, industrial companies are exploring smarter and more efficient methods of production planning and scheduling. An optimal production plan requires meeting production requirements and constraints at the lowest cost. Opcenter APS (previously known as "Preactor APS") was developed specifically to meet this requirement as an advanced planning and scheduling software solution for generating achievable production plans by balancing demand and capacity. Opcenter APS provides users with professional production data management methods and simple and efficient production scheduling rules.

SUMMARY OF THE INVENTION

In view of this, the embodiments of the present invention propose a simulation-based closed-loop APS scheduling optimization method on the one hand, and a simulation-based closed-loop APS scheduling optimization system and a computer-readable storage medium on the other hand, so as to provide more feasible scheduling scheme.

A simulation-based closed-loop APS scheduling optimization method proposed in the embodiment of the present invention includes: determining an APS rule selected for scheduling; determining a weight configuration set for the selected APS rule; according to the weight configuration of the APS rule, Generate a scheduling Gantt chart; export the order sequence corresponding to the scheduling Gantt chart; load the order sequence and the obtained preset simulation model configuration data into the basic simulation model of the factory design simulation software to obtain scheduling simulation model; run the scheduling simulation model, and evaluate to obtain key performance indicator data; export the key performance indicator data as the simulation result; determine whether the simulation result meets the needs, and if so, output the corresponding scheduling Gantt chart ; otherwise, return to the step of selecting APS rules for scheduling.

In one embodiment, the determining the weight configuration set for the selected APS rule includes: using a comprehensive particle swarm optimization algorithm to set the weight configuration for the selected APS rule; the comprehensive particle swarm optimization algorithm is in the basic particle swarm optimization algorithm. Improvement is made on the basis of the iterative optimization process. In the iterative optimization process, the fitness value is calculated according to the key performance indicator data to determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm. The optimal position of the particle itself is evaluated and processed, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position of the particle swarm, dynamically adjust the inertia weight value, and finally update the next-generation particle swarm.

In one embodiment, the setting of the weight configuration for the selected APS rule by using the comprehensive particle swarm optimization algorithm includes: assigning an initial weight to the selected APS rule, and performing particle encoding on the initial weight; Initialize the weight of the APS to obtain the current particle swarm; decode the current particle swarm to obtain the current weight configuration of the APS rule, and determine the current weight configuration as the weight configuration set for the selected APS rule; After executing the step of running the scheduling simulation model and evaluating and obtaining key performance indicator data, the step further includes: acquiring the key performance indicator data, and calculating a fitness value according to the key performance indicator data to determine the particle itself experienced The optimal position experienced by the particle swarm and the optimal position experienced by the particle swarm; determine whether the maximum number of iterations has been reached? When the maximum number of iterations is not reached, the simulated annealing algorithm is used to evaluate the optimal position experienced by the particle itself, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position experienced by the particle swarm; A new generation of the current particle swarm, and return to perform the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule; when the maximum number of iterations is reached, the optimal The particle swarm corresponding to the position is used as the current example swarm, and the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule is returned to execute.

In one embodiment, the determining the APS rule to be selected for the schedule includes: selecting the rule required for the current schedule based on the existing rules and/or added custom rules of the advanced planning and scheduling software.

The simulation-based closed-loop APS scheduling optimization system proposed in the embodiment of the present invention includes: at least one memory and at least one processor, wherein: the at least one memory is used to store a computer program; the at least one processor is used to call the A computer program stored in at least one memory causes the device to perform corresponding operations, the operations comprising: determining an APS rule selected for scheduling; determining a weight configuration set for the selected APS rule; weighting according to the APS rule configure and generate a scheduling Gantt chart; export the order sequence corresponding to the scheduling Gantt chart; load the order sequence and the acquired preset simulation model configuration data into the basic simulation model of the factory design simulation software to obtain the scheduling run the scheduling simulation model, and evaluate to obtain key performance indicator data; export the key performance indicator data as the simulation result; judge whether the simulation result meets the needs, and if so, output the corresponding scheduling plan special graph; otherwise, go back to executing the step of selecting APS rules for the schedule.

The simulation-based closed-loop APS scheduling optimization system proposed in the embodiment of the present invention includes: an advanced planning and scheduling software module; a factory design simulation simulation software module; and a simulation-based scheduling module, which is constructed as a component object model COM component, Integrate into the advanced planning and scheduling software module through a COM interface, connect to the plant design simulation software module through the COM interface of the plant design simulation software module, and perform the following operations: determine the APS rule; determine the weight configuration of the selected APS rule; generate a scheduling Gantt chart according to the weight configuration of the APS rule; derive the order sequence corresponding to the scheduling Gantt chart; The simulation model configuration data is loaded into the basic simulation model of the factory design simulation simulation software module to obtain a scheduling simulation model; run the scheduling simulation model, and evaluate to obtain key performance indicator data; use the key performance indicator data as the simulation Deriving the result; judging whether the simulation result meets the requirement, if yes, outputting the corresponding scheduling Gantt chart; otherwise, returning to the step of selecting an APS rule for scheduling.

In one embodiment, the determining the weight configuration of the selected APS rule includes: using a comprehensive particle swarm optimization algorithm to set a weight configuration for the selected APS rule; the comprehensive particle swarm optimization algorithm is based on the basic particle swarm optimization algorithm. In the iterative optimization process, the fitness value is calculated according to the key performance indicator data to determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm, and the simulated annealing algorithm is used for each obtained particle. The best position itself is evaluated and processed, and the mutation operation of the genetic algorithm is used to evaluate and update the best position of the particle swarm, dynamically adjust the inertia weight value, and finally update the next generation of particle swarm.

In one embodiment, the simulation-based scheduling module includes: a scheduling sub-module and a simulation sub-module; wherein the scheduling sub-module is used to determine the APS rule selected for scheduling, and determine the weight of the selected APS rule configuration; according to the weight configuration of the APS rules, generate a scheduling Gantt chart; export the order sequence corresponding to the scheduling Gantt chart to the database; and obtain the simulation results based on the order sequence from the database; judge the Whether the simulation result meets the requirements, if yes, output the corresponding scheduling Gantt chart; otherwise, return to execute the operation of selecting APS rules for scheduling; the simulation submodule is used to load the order sequence from the database and preset simulation model configuration data, and load the ordered sequence and the simulation model configuration data into the basic simulation model of the factory design simulation software to obtain a scheduling simulation model; run the scheduling simulation model, and evaluating to obtain key performance indicator data; and storing the key performance indicator data in the database as a simulation result.

In one embodiment, the simulation-based scheduling module further includes: an optimization algorithm sub-module; configured to set a weight configuration for the selected APS rule by using a comprehensive particle swarm optimization algorithm, and provide the weight configuration to the scheduling sub-module Module; the integrated particle swarm optimization algorithm is improved on the basis of the basic particle swarm optimization algorithm, and in the iterative optimization process, the fitness value is calculated according to the key performance indicator data to determine the optimal position experienced by the particle itself and the particle swarm experience. The optimal position of the particle swarm is evaluated and processed by the simulated annealing algorithm, and the optimal position of the particle swarm is evaluated and updated by the mutation operation of the genetic algorithm, and the inertia weight value is dynamically adjusted, and the final update is generated. Next-generation particle populations.

In one embodiment, the optimization algorithm sub-module assigns initial weights to the selected APS rules, and performs particle encoding on the initial weights; initializes the weights after particle encoding to obtain the current particle swarm; The current particle swarm is decoded to obtain the current weight configuration of the APS rule, and the current weight configuration is provided to the scheduling sub-module; the key performance indicator data evaluated by the simulation sub-module is obtained, and The performance index data calculates the fitness value to determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm; determine whether the maximum number of iterations has been reached? When the maximum number of iterations is not reached, the simulated annealing algorithm is used to evaluate the optimal position experienced by the particle itself, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position experienced by the particle swarm; A new generation of the current particle swarm, and return to perform the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule; when the maximum number of iterations is reached, the optimal The particle swarm corresponding to the position is used as the current example swarm, and the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule is returned to execute.

In one embodiment, the simulation-based scheduling module further includes: an APS rule configuration module for determining the rules required for the current scheduling based on the existing rules and/or added custom rules of the advanced planning and scheduling software.

The computer-readable storage medium proposed in the embodiment of the present invention stores a computer program thereon; the computer program can be executed by a processor and implements the above simulation-based closed-loop APS scheduling optimization method.

It can be seen from the above scheme that, since the advanced planning and scheduling software and the factory design simulation software are integrated in the embodiment of the present invention, the scheduling scheme generated based on the advanced planning and scheduling software can be simulated by the factory design simulation software. Create a corresponding scheduling simulation model, and run the scheduling simulation model to obtain the KPI data of the simulation results. Based on the KPI data, it can be judged whether the corresponding scheduling scheme meets the requirements, so that a more feasible scheduling scheme can be provided.

Further, by using the comprehensive particle swarm optimization algorithm to automatically assign the weights to the APS rules selected by the scheduling, those who face new production scenarios or changing production requirements do not know how to choose the production plan with the weight of the APS rules. Personnel can complete the schedule accurately and efficiently.

Description of drawings

The above-mentioned and other features and advantages of the present invention will be more apparent to those of ordinary skill in the art by describing the preferred embodiments of the present invention in detail below with reference to the accompanying drawings, in which:

FIG. 1 is an exemplary flowchart of a simulation-based closed-loop APS scheduling optimization method in an embodiment of the present invention.

FIG. 2 is a schematic diagram of a framework structure of a simulation-based closed-loop APS scheduling optimization system according to an embodiment of the present invention.

FIG. 3 is a schematic flowchart of a simulation-based closed-loop APS scheduling optimization method based on the framework shown in FIG. 2 in an embodiment of the present invention.

FIG. 4 is a schematic diagram of a function dialog of the SSM in an example of the present invention.

FIG. 5 is a schematic diagram of the configuration of the APS rule base in an example of the present invention.

FIG. 6 is a schematic diagram of a process of creating a simulation model in an example of the present invention.

FIG. 7 is a schematic diagram of an automatic optimization process of APS rule weight assignment in an example of the present invention.

FIG. 8 is an exemplary structural diagram of a simulation-based closed-loop APS scheduling optimization system in an embodiment of the present invention.

Among them, the reference numerals are as follows:

Detailed ways

In the embodiment of the present invention, considering that up to now, advanced planning and scheduling software such as Opcenter APS, like other industrial APS software, does not have an optimization algorithm for scheduling, but provides a set of APS rules to generate a scheduling Gantt chart. It is necessary to manually adjust and optimize the results according to personal experience, which makes Opcenter APS still have some shortcomings, especially in some complex production scenarios, there are the following two problems: 1. It is difficult to verify and evaluate the planning and scheduling results given by Opcenter APS 2. There is no mature Opcenter APS algorithm to optimize the rule-based scheduling scheme.

The first problem above makes the scheduling results of Opcenter APS infeasible or poorly executed in the actual production process, and rescheduling always costs extra. The second problem comes from the rule-based scheduling algorithm provided by Opcenter APS. APS rules enable very fast scheduling, but are not user-oriented. As a result, users often cannot find a good way to assess their quality, see if the schedule is in line with production requirements, or if a better solution can be found. Therefore, Opcenter APS needs a feasible method to optimize the APS rules and verify the scheduling results for different user target requirements.

In response to the first question, some studies are currently considering providing a calculator for generating key performance indicators (KPI, Key Performance Indicators) displayed through various charts or reports based on a mathematical model based on the current scheduling results in the Gantt chart. )Statistics. Among them, the charts may include "utilization" charts representing the utilization of production resources, etc., and the reports may include order KPIs related to time and cost. Using KPI statistics, users can quickly evaluate scheduling results and make adjustments.

For the second problem, in some academic studies, various heuristic algorithms, such as genetic algorithm (GA), particle swarm optimization (PSO), ant colony algorithm (AG) and simulated annealing algorithm (SA), are considered to solve the most Optimal order scheduling sequence. But all these smart algorithms require a lot of computing resources and time. In some complex production scenarios, this intelligent algorithm cannot find the global optimal solution, but can only find the local optimal solution. On the other hand, optimization algorithms have to be developed one by one. Therefore, it is difficult to apply different scenarios in actual production.

Therefore, in this embodiment of the present invention, it is considered to provide a simulation-based closed-loop APS scheduling optimization solution. This solution considers replacing the data model with the model in plant design simulation software such as Plant Simulation to simulate planned production and calculate KPI statistics; and consider developing an optimization algorithm in Opcenter APS to optimize APS rules.

Specifically, a simulation-based Scheduling Module (SSM) can be provided for integrating advanced planning and scheduling software such as Opcenter APS and plant design simulation software such as Plant Simulation to realize a simulation-based closed-loop APS scheduling optimization scheme. Provide the scheduling scheme based on advanced planning and scheduling software such as Opcenter APS to the simulation model generated based on plant design simulation software such as Plant Simulation for verification, obtain the corresponding KPI data according to the simulation results, and judge whether the scheduling scheme meets the requirements based on the KPI data. Require. For example, a simulation-based closed-loop APS scheduling optimization method can be shown in Figure 1, which includes the following steps:

Step S101, determining the APS rule selected for the schedule.

Step S102, determining the weight configuration set for the selected APS rule.

Step S103, generating a scheduling Gantt chart according to the weight configuration of the APS rule.

Step S104, derive the order sequence corresponding to the scheduling Gantt chart.

Step S105, load the order sequence and the acquired preset simulation model configuration data into the basic simulation model of the factory design simulation software to obtain a scheduling simulation model.

Step S106, running the scheduling simulation model, and evaluating to obtain key performance indicator data.

Step S107, exporting the key performance indicator data as a simulation result.

In step S108, it is judged whether the simulation result meets the requirements, if yes, step S109 is performed; otherwise, step S101 is returned to.

Step S109, output the corresponding scheduling Gantt chart.

Further, when the scheduling plan is formulated based on the advanced planning and scheduling software, the optimal iterative algorithm can be used to find the optimal weight of each APS rule in the schedule, and the KPI data obtained by simulation can be optimized accordingly when finding the final weight. calculate. For example, a comprehensive particle swarm optimization algorithm can be provided that sets a weight configuration for selected APS rules. The comprehensive particle swarm optimization algorithm is improved on the basis of the basic particle swarm optimization algorithm. In the iterative optimization process, the fitness value is calculated according to the key performance indicator data to determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm. The optimal position of the particle swarm is evaluated and processed by the simulated annealing algorithm, and the optimal position of the particle swarm is evaluated and updated by the mutation operation of the genetic algorithm, and the inertia weight value is dynamically adjusted, and finally the next generation particle is generated. population.

In order to make the objectives, technical solutions and advantages of the present invention clearer, the following examples are used to further describe the present invention in detail.

FIG. 2 is a schematic diagram of a framework structure of a simulation-based closed-loop APS scheduling optimization system according to an embodiment of the present invention. In this embodiment, a case where the advanced planning and scheduling software is Opcenter APS and the factory design simulation software is Plant Simulation is used as an example for description. As shown in Figure 2, the framework structure is divided into three layers: application layer 1, service layer 2 and data layer 3.

Application Layer 1 represents the integration of Opcenter APS 10 and Plant Simulation20. Specifically, in the embodiment of the present invention, an SSM 30 is provided, and the SSM 30 is constructed as a Component Object Model (COM, Component Object Model) component, which is integrated into the Opcenter APS 10 through the COM interface, and is integrated into the Opcenter APS 10 through the COM interface of the Plant Simulation20. The interface is connected to Plant Simulation20, thereby obtaining the simulation-based closed-loop APS scheduling optimization system in the embodiment of the present invention.

The service layer 2 includes three parts: the scheduling sub-module 31 formed by the SSM 30 using the APS framework and the application programming interface (API), the SSM 30 using the framework and API of the Plant Simulation 20 to form a simulation sub-module including a high-precision simulation model 321. 32 and the optimization algorithm sub-module 33 provided by the SSM 30.

The scheduling sub-module 31 includes an APS rule base 311 and a scheduler 312 . The APS rule base 311 includes various APS rules, which not only include the built-in rules of Opcenter APS, but also include user-defined rules. The scheduler 312 is the scheduling engine that performs the scheduling process.

In the service layer 2, the optimization algorithm sub-module 33 is used to schedule the corresponding rules currently selected in the APS rule base 311, and assign an initial weight to each rule, perform an encoding operation 331 on the assigned initial weight, and generate an iterative Initial weight assignments for particle populations. In order to verify and evaluate each weight assignment, the weight assignment particles are decoded 332 and provided to the scheduling sub-module 31 for simulation verification by the simulation sub-module 32, and the KPI data reported by the simulation sub-module 32 are received 322 , iteratively optimize each weight assignment according to the KPI data, perform decoding operation 332 on the final weight assignment after the iteration, and provide the decoded weight assignment to the scheduling sub-module 31 to generate an optimal scheduling scheme. The data required for operations in this process comes from data layer 3.

The data layer 3 includes the data sources of all the modules of the above-mentioned service layer 2. Basically, it is a SQL Server database 40 serving Opcenter APS 10, which contains resource data 41, product data 42, inventory data 43, order data 44, etc., which are necessary for scheduling and simulation. On this basis, the database also stores scheduling result data (such as order sequence, etc.) 45 and simulation model configuration data 46, which are required for simulation.

In this embodiment, both the scheduling input/output data and the data required for the simulation are stored in the same database, and a standard data table template can be used. From order scheduling to simulated production, Data Layer 3 can also be applied to possible scenarios such as MRP (Material Requirements Planning), MPS (Master Production Planning), SCM (Supply Chain Management).

3 is a schematic flowchart of a simulation-based closed-loop APS scheduling optimization method based on the framework shown in FIG. 2 in an embodiment of the present invention. As shown in FIG. 3 , the method mainly involves the interaction between the scheduling sub-module 31 , the simulation sub-module 32 and the optimization algorithm sub-module 33 .

Specifically, the following operations may be included on the scheduling sub-module 31 side:

Step S401, import scheduling data.

In this step, the scheduling data is the data input required for scheduling, such as order, resource, process and other related data.

Step S402, setting an optimization target.

In this step, the optimization goal is also the scheduling goal, for example, whether all orders are delivered or the finished product inventory is the lowest, and/or the work-in-process inventory is the lowest, and/or some orders are processed in advance, and/or the inventory is within a range internal, and/or lowest cost, etc.

Step S403, determining the APS rule selected for the schedule.

In this step, the user can select the required APS rule, or select the required APS rule according to the pre-configuration. In a specific implementation, it can be shown in FIG. 4 , which is a schematic diagram of a function dialog box of the SSM 30 in an example of the present invention. It can be seen that an APS rule configuration module 34 may be specifically included on the SSM 30 side. Opcenter APS 10 provides some built-in APS rules, such as by the shortest delivery time (EDD), shortest processing time (SPT), lowest critical ratio (CR), first in first out (FIFO) and other rules. Furthermore, in addition to the built-in rules, the APS rule base further extends the rules such as Shortest Relaxation Time (STR), Last In First Out (LIFO), Least Number of Operations (LOPNR), etc. The APS rule configuration module 34 presents these rules to the user, and the user selects the required rules. Of course, the rules library is also available for users to define and add custom rules by setting properties for each production order and adding those properties to the library as new rules for order sorting. FIG. 5 shows a schematic diagram of the configuration of the APS rule base in an example. As shown in Figure 5, the right side is a part of the rules in the presented Opcenter APS rule base, and the user can perform three operations through the APS rule configuration module 34: the first operation 341: add a custom rule; the second operation 342: select the selected For the required rules, EDD, LIFO, and LOPNR are selected on the right side of FIG. 5 ; the third operation 343 : assign weights to the selected rules, such as 1, 2, and 3 in the boxes on the right side of FIG. 5 .

Among them, regarding the third operation 341 of assigning weights to the selected rules, there may be two implementation solutions in this embodiment. One is manual scheduling, that is, the user manually assigns weights to the selected APS rules, and then invokes the simulation model to verify the scheduling scheme. This approach is suitable for experienced planners and those who need quick planning and verification. If the manual scheduling scheme is adopted, step S404 is performed after this step, and if automatic scheduling is selected, step S405 is performed.

Step S404, accept the weight configuration of the APS rule by the user.

Step S405, accept the weight configuration of the APS rule by the optimization algorithm sub-module 33.

Wherein, the weight configuration of the APS rules by the optimization algorithm submodule 33 refers to that the SSM 30 utilizes the optimization algorithm submodule 33 to automatically configure the weights of the APS rules, and continuously searches for the best APS scheduling sequence for the configured weights through scheduling and simulation. Autoscheduling is designed for planners who are faced with new production scenarios or changing goals and do not know how to choose APS rule weights. Therefore, simulating the closed-loop scheduling optimization scheme helps users to verify the scheduling results and automatically optimize the APS rules. For the specific process, please refer to steps 501 to 509 below.

In step S406, a scheduling scheme is generated according to the weight configuration of the APS rule, that is, a scheduling Gantt chart is obtained. The processing sequence of each order and its corresponding processing station and processing time are reflected in the scheduling Gantt chart.

In this step, the total score of each order can be calculated according to the following formula (1) according to the weight configuration of the APS rules.

S _sum = (V _fcfs ×W _fcfs +V _edd ×W _edd +V _spt ×W _spt +V _str ×W _str )/(W _fcfs +W _edd +W _spt +W _str ) (1)

S _sum : the total score of each order;

V _fcfs : the initial priority score of each order, the value of the first order is 1, the value of the first order is 2, and so on;

V _edd : the initial delivery time score of each order, the value of the order with the earliest delivery date is 1, the value of the order with the second earliest delivery date is 2, and so on;

V _spt : the initial processing time score of each order, the value of the order with the shortest processing time is 1, the value of the order with the second shortest processing time is 2, and so on;

V _str : The initial slack time score for each order, the order with the shortest remaining time has a value of 1, the order with the second shortest remaining time has a value of 2, and so on.

W _fcfs , W _edd , W _spt , and W _str are weight values.

The score of each order determines the entire order sequence and ultimately the scheduling result.

Step S407, based on the scheduling scheme, derive the corresponding order sequence, and store the order sequence in a database (DB, Database).

Step S411, extract the simulation result from the database.

Step S412, it is judged whether the simulation result satisfies the requirement, if yes, execute step S413; otherwise, return to execute step S403.

Step S413, output the corresponding scheduling scheme.

It can be seen that the scheduling sub-module 31 is mainly used to determine the APS rule selected for scheduling, and determine the weight configuration of the selected APS rule; generate a scheduling Gantt chart according to the weight configuration of the APS rule; derive the corresponding scheduling Gantt The order sequence of the special diagram is stored in the database; and the simulation result based on the order sequence is obtained from the database; it is judged whether the simulation result meets the needs, and if so, the corresponding scheduling Gantt chart is output; Describes the operation of scheduling the selection of APS rules.

The following operations may be included on the side of the simulation sub-module 32:

In step S408, the simulation sub-module 32 loads simulation-related data and creates a scheduling simulation model. The loading of the simulation-related data includes: step S4081, loading the simulation model configuration data; step S4082, loading the simulation model; and step S4083, loading the order sequence.

In this step S408, the simulation model can be quickly created as shown in FIG. 6 . Various methods may be used through the COM interface, and the specific process may include: First, the SSM 30 loads the basic simulation model through the LoadModel() method. There are a series of SimTalk methods and ODBC connection objects in the basic simulation model, which are remotely controlled by the ExecuteSimTalk() method, which can read data from the Opcenter APS database, and then automatically create and run the relevant production line simulation model. In addition, the order sequence generated by the scheduling sub-module 31 is also loaded into the current production line simulation model and distributed to each production station. In order to calibrate the model, simulation model configuration data is also loaded in this way, which can be entered by the user from the SSM 30 integrated into Opcenter APS 10 and stored in the database. Simulation model configuration data may include settings for workstation processing time, workstation mean time to failure (MTTF), yield, logistics rules, and more. This results in a high-precision model, and each time by changing production-related data in Opcenter APS10, a new model can be quickly created to run the simulation. Then after the simulation triggers the SimulationFinished event, the SSM finally receives the KPI data through the GetValue() method. The first interface 61 and the second interface 62 in FIG. 6 are only an example carried by the creation of a schematic representation model, which are not used to limit the actual application interface, and do not affect the implementation of the technical solution of the present invention, so they are not used to limit the actual application interface. The specific content above is obfuscated.

Step S409, running the scheduling simulation model and evaluating the simulated KPI data.

In this embodiment, if the weights of the aforementioned APS rules are automatically configured by the optimization algorithm sub-module 33 , the KPI data obtained in this step may be fed back to the optimization algorithm sub-module 33 .

In step S410, the KPI data is exported and stored in the database DB as the simulation result.

It can be seen that the simulation sub-module 32 is mainly used to load the order sequence and the preset simulation model configuration data from the database, and load the order sequence and the simulation model configuration data into the basic simulation model of Plant Simulation , obtain a scheduling simulation model; run the scheduling simulation model, and evaluate to obtain key performance indicator data; store the key performance indicator data as a simulation result in the database.

The following operations may be included on the side of the optimization algorithm sub-module 33:

In this embodiment, considering that the scheduling optimization problem is based on discrete scenarios, and in this application, it is considered to guide the weight configuration of APS rules based on KPI results. The optimization algorithm of the group optimization algorithm (SG-DPSO). The SG-DPSO is iteratively optimized based on the particle swarm optimization (PSO, Particle swarm optimization), and in the iterative optimization process, the SA algorithm and the GA algorithm are combined to complete the optimization of the weight configuration. Among them, the initialization of PSO is a group of random particles (random solutions), and then the optimal solution is found through iteration. In each iteration, the particle updates itself by tracking two "extremes", namely the optimal position Pbest experienced by the particle itself and the optimal position Gbest experienced by the particle swarm. After finding these two optimal solutions, the particle updates its velocity and position. The reason for using the PSO algorithm is that the result of each iteration can guide the particle's fitness value to get better by updating the particle's velocity and position. At the same time, the SA algorithm is used to process the current optimal particle position, and the mutation operation of the GA algorithm is used to process and update the global optimal position of the particle population, thereby avoiding the generation of local optimal solutions. Furthermore, fitness values are calculated from the KPI data to determine Pbest and Gbest. The specific implementation may include the following steps:

Step S501: Determine the APS rule selected by the scheduling sub-module 31 side through the COM interface, assign an initial weight to the APS rule, and perform particle coding on the initial weight.

For example, FIG. 7 shows a schematic diagram of an automatic optimization process of APS rule weight assignment. As shown in Figure 7, the APS rule combination includes: the selected four kinds of rules: EDD, FIFO, STR and SPT, a correspondingly encoded possible particle should be like "1342", and the number of each position represents each The weight value of the rule, the maximum value indicates that the relevant rule has the highest weight. After decoding, the scheduling sub-module 31 can calculate the total score of each production order according to the weight values of different APS rules.

Step S502: Initialize the weight after particle encoding to obtain the current particle swarm.

Step S503, decode the current particle swarm to obtain the current weight configuration of the APS rule, and provide the current weight configuration of the APS rule to the scheduling sub-module 31 through the COM interface, and the scheduling sub-module 31 will perform the step S404. Accept the weight configuration of the APS rule by the optimization algorithm sub-module 33 .

In step S504, the KPI data fed back by the simulation sub-module 32 after step S408 is received through the COM interface, and the fitness value is calculated according to the KPI data to determine Pbest and Gbest.

Step S505, determine whether the maximum number of iterations has been reached? If yes, go to step S509; otherwise, go to step S506, and adjust the inertia weight value updated by the particle position according to the number of iterations.

Step S506, using a simulated annealing algorithm to evaluate Pbest.

Step S507, using the mutation operation of the genetic algorithm to evaluate and update the Gbest.

In step S508, a new generation of the current particle swarm is generated, and then step S503 is performed.

In step S509, the particle swarm corresponding to Gbest is selected as the current example swarm, and then step S503 is executed.

It can be seen that the optimization algorithm sub-module 33 is mainly used to set the weight configuration for the selected APS rule by using the comprehensive particle swarm optimization algorithm, and provide the weight configuration to the scheduling sub-module; the comprehensive particle swarm optimization algorithm is used in the basic particle swarm optimization Improvements are made on the basis of the optimization algorithm. In the iterative optimization process, the fitness value is calculated according to the key performance indicator data to determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm. The optimal position of the particle itself is evaluated and processed, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position of the particle swarm, dynamically adjust the inertia weight value, and finally update the next-generation particle swarm.

It can be seen that, as shown in FIGS. 2 to 7 , a simulation-based closed-loop APS scheduling optimization system in an embodiment of the present invention may include: an advanced planning and scheduling software module, a factory design simulation software module, and a simulation-based scheduling module 30, which is constructed as a component object model COM component, integrated into the high-level planning and scheduling software module through a COM interface, and connected to the plant design simulation software module through the COM interface of the plant design simulation software module module, and perform the following operations: determine the APS rule selected for scheduling; determine the weight configuration of the selected APS rule; generate a scheduling Gantt chart according to the weight configuration of the APS rule; export the corresponding scheduling Gantt chart order sequence; load the order sequence and the acquired preset simulation model configuration data into the basic simulation model of the factory design simulation simulation software module to obtain a scheduling simulation model; run the scheduling simulation model, and evaluate to obtain the key performance indicator data; exporting the key performance indicator data as the simulation result; judging whether the simulation result meets the requirements, if so, output the corresponding scheduling Gantt chart; otherwise, return to execute the described selection APS rule for scheduling step.

FIG. 8 is a schematic structural diagram of another simulation-based closed-loop APS scheduling optimization system in an embodiment of the present invention, and the apparatus can be used to implement the methods shown in FIG. 1 and FIG. 3 , or implement the system shown in FIG. 2 . As shown in FIG. 8 , the system may include: at least one memory 81 , at least one processor 82 and at least one display 83 . In addition, some other components, such as communication ports, etc., may also be included. These components communicate via bus 84 .

Among them, at least one memory 81 is used to store computer programs. In one embodiment, the computer program can be understood as including each module of the simulation-based closed-loop APS scheduling optimization system shown in FIG. 2 . In addition, the at least one memory 81 may also store an operating system and the like. Operating systems include but are not limited to: Android operating system, Symbian operating system, Windows operating system, Linux operating system, and so on.

At least one display 83 is used to display a scheduling Gantt chart or the like.

At least one processor 82 is configured to call a computer program stored in at least one memory 81 to execute the simulation-based closed-loop APS scheduling optimization method described in the embodiments of the present invention.

Specifically, at least one processor 82 is configured to invoke a computer program stored in at least one memory 81 to cause the apparatus to perform corresponding operations. The operations may include: determining an APS rule selected for scheduling; determining a weight configuration set for the selected APS rule; generating a scheduling Gantt chart according to the weight configuration of the APS rule; exporting a corresponding scheduling Gantt chart order sequence; load the order sequence and the obtained preset simulation model configuration data into the basic simulation model of Plant Simulation to obtain a scheduling simulation model; run the scheduling simulation model, and evaluate and obtain KPI data; The KPI data is derived as the simulation result; it is judged whether the simulation result meets the requirements, and if so, the corresponding scheduling Gantt chart is output; otherwise, the step of selecting APS rules for scheduling is returned to.

In one embodiment, the setting of the weight configuration for the selected APS rule by using the comprehensive particle swarm optimization algorithm includes: assigning an initial weight to the selected APS rule, and performing particle encoding on the initial weight; Initialize the weight of the APS to obtain the current particle swarm; decode the current particle swarm to obtain the current weight configuration of the APS rule, and determine the current weight configuration as the weight configuration set for the selected APS rule; After executing the step of running the scheduling simulation model and evaluating and obtaining key performance indicator data, the step further includes: acquiring the key performance indicator data, and calculating a fitness value according to the key performance indicator data to determine the particle itself experienced The optimal position experienced by the particle swarm and the optimal position experienced by the particle swarm; determine whether the maximum number of iterations has been reached? When the maximum number of iterations is not reached, the simulated annealing algorithm is used to evaluate the optimal position experienced by the particle itself, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position experienced by the particle swarm; A new generation of the current particle swarm, and return to perform the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule; when the maximum number of iterations is reached, the optimal The particle swarm corresponding to the position is taken as the current example swarm, and the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule is returned to be executed.

It can be seen from the above scheme that since Opcenter APS and Plant Simulation are integrated in the embodiment of the present invention, the scheduling scheme generated based on Opcenter APS can create a corresponding scheduling simulation model through Plant Simulation, and by running the scheduling scheme The process simulation model can obtain the KPI data of the simulation results. Based on the KPI data, it can be judged whether the corresponding scheduling scheme meets the requirements, so that a more feasible scheduling scheme can be provided.

Further, by using the comprehensive particle swarm optimization algorithm to automatically assign the weights to the APS rules selected by the scheduling, those who face new production scenarios or changing production requirements do not know how to choose the production plan with the weight of the APS rules. Personnel can complete the schedule.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the present invention. within the scope of protection.

Claims

The simulation-based closed-loop advanced planning and scheduling software (APS) scheduling optimization method is characterized by comprising:

Determine the APS rule selected for the schedule (S101);

Determine the weight configuration set for the selected APS rule (S102);

According to the weight configuration of the APS rule, a scheduling Gantt chart is generated (S103);

Deriving an order sequence corresponding to the scheduling Gantt chart (S104);

Loading the order sequence and the acquired preset simulation model configuration data into the basic simulation model of the factory design simulation software to obtain a scheduling simulation model (S105);

Running the scheduling simulation model, and evaluating to obtain key performance indicator data (S106);

Exporting the key performance indicator data as a simulation result (S107);

It is judged whether the simulation result meets the requirements (S108), and if so, the corresponding scheduling Gantt chart is output (S109); otherwise, it returns to execute the step of selecting an APS rule for scheduling.
The simulation-based closed-loop APS scheduling optimization method according to claim 1, wherein the determining the weight configuration (S102) set for the selected APS rule comprises:

A comprehensive particle swarm optimization algorithm is used to set the weight configuration for the selected APS rules; the comprehensive particle swarm optimization algorithm is improved on the basis of the basic particle swarm optimization algorithm, and it calculates the fitness value according to the key performance indicator data in the iterative optimization process Determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm, and use the simulated annealing algorithm to evaluate the optimal position of the particle itself, and use the mutation operation of the genetic algorithm to optimize the particle swarm. The position is evaluated and updated, the inertia weight value is dynamically adjusted, and finally the next generation particle population is generated by the update.
The simulation-based closed-loop APS scheduling optimization method according to claim 2, wherein the setting of the weight configuration for the selected APS rules by using a comprehensive particle swarm optimization algorithm comprises:

assigning initial weights to the selected APS rules, and performing particle encoding on the initial weights;

Initialize the weights after particle encoding to get the current particle swarm;

Decoding the current particle swarm to obtain the current weight configuration of the APS rule, and determining the current weight configuration as the weight configuration set for the selected APS rule;

After executing the step of running the scheduling simulation model and evaluating and obtaining key performance indicator data (S106), the method further includes: acquiring the key performance indicator data, and calculating a fitness value according to the key performance indicator data to obtain the key performance indicator data. Determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm;

Determine whether the maximum number of iterations has been reached? When the maximum number of iterations is not reached, the simulated annealing algorithm is used to evaluate the optimal position experienced by the particle itself, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position experienced by the particle swarm; A new generation of the current particle swarm, and returns to perform the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule;

When the maximum number of iterations is reached, take the particle swarm corresponding to the optimal position experienced by the particle swarm as the current example swarm, and return to perform the decoding of the current particle swarm to obtain the current weight configuration of the APS rule operation.
The simulation-based closed-loop APS scheduling optimization method according to any one of claims 1 to 3, wherein the determining the APS rule (S101) selected for scheduling comprises: based on existing advanced planning and scheduling software Rules and/or Added Custom Rules Select the rules required for the current schedule.
The simulation-based closed-loop advanced planning and scheduling software (APS) scheduling optimization system is characterized by comprising: at least one memory (81) and at least one processor (82), wherein:

the at least one memory (81) is used to store computer programs;

The at least one processor (82) is configured to invoke a computer program stored in the at least one memory (81) to cause the apparatus to perform corresponding operations, the operations comprising:

Determine the APS rules selected for the schedule;

Determine the weight configuration set for the selected APS rule;

generating a scheduling Gantt chart according to the weight configuration of the APS rule;

Deriving an order sequence corresponding to the scheduling Gantt chart;

Loading the order sequence and the obtained preset simulation model configuration data into the basic simulation model of the factory design simulation software to obtain a scheduling simulation model;

Running the scheduling simulation model and evaluating the key performance indicator data;

exporting the key performance indicator data as simulation results;

It is judged whether the simulation result satisfies the requirement, and if so, the corresponding scheduling Gantt chart is output; otherwise, it returns to execute the step of selecting an APS rule for scheduling.
The simulation-based closed-loop APS scheduling optimization system according to claim 5, wherein,

The weight configuration determined to be set for the selected APS rule includes:

A comprehensive particle swarm optimization algorithm is used to set the weight configuration for the selected APS rules; the comprehensive particle swarm optimization algorithm is improved on the basis of the basic particle swarm optimization algorithm, and it calculates the fitness value according to the key performance indicator data in the iterative optimization process Determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm, and use the simulated annealing algorithm to evaluate the optimal position of the particle itself, and use the mutation operation of the genetic algorithm to optimize the particle swarm. The position is evaluated and updated, the inertia weight value is dynamically adjusted, and finally the next generation particle population is generated by the update.
The simulation-based closed-loop APS scheduling optimization system according to claim 6, characterized in that, the use of a comprehensive particle swarm optimization algorithm to set a weight configuration for the selected APS rules comprises:

assigning initial weights to the selected APS rules, and performing particle encoding on the initial weights;

Initialize the weights after particle encoding to get the current particle swarm;

Decoding the current particle swarm to obtain the current weight configuration of the APS rule, and determining the current weight configuration as the weight configuration set for the selected APS rule;

After executing the step of running the scheduling simulation model and evaluating and obtaining key performance indicator data, the method further includes: acquiring the key performance indicator data, and calculating a fitness value according to the key performance indicator data to determine the particle itself The optimal position experienced and the optimal position experienced by the particle swarm;

Determine whether the maximum number of iterations has been reached? When the maximum number of iterations is not reached, the simulated annealing algorithm is used to evaluate the optimal position experienced by the particle itself, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position experienced by the particle swarm; A new generation of the current particle swarm, and returns to perform the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule;

When the maximum number of iterations is reached, take the particle swarm corresponding to the optimal position experienced by the particle swarm as the current example swarm, and return to perform the decoding of the current particle swarm to obtain the current weight configuration of the APS rule operation.
The simulation-based closed-loop APS scheduling optimization system according to any one of claims 5 to 7, wherein the determining the APS rule selected for scheduling comprises: based on the existing rules of advanced planning and scheduling software and/or Or add custom rules to select the rules required for the current schedule.
The simulation-based closed-loop advanced planning and scheduling software (APS) scheduling optimization system is characterized in that it includes:

Advanced planning and scheduling software modules;

plant design simulation software modules; and

A simulation-based scheduling module (30), which is structured as a component object model COM component, is integrated into the high-level planning and scheduling software module through a COM interface, and is connected to the plant design simulation software module through a COM interface of the simulation software module The factory design simulates the simulation software module, and performs the following operations:

Determine the APS rules selected for the schedule;

Determine the weight configuration of the selected APS rule;

generating a scheduling Gantt chart according to the weight configuration of the APS rule;

Deriving an order sequence corresponding to the scheduling Gantt chart;

Loading the order sequence and the acquired preset simulation model configuration data into the basic simulation model of the factory design simulation simulation software module to obtain a scheduling simulation model;

Running the scheduling simulation model and evaluating the key performance indicator data;

exporting the key performance indicator data as simulation results;

It is judged whether the simulation result satisfies the requirement, and if so, the corresponding scheduling Gantt chart is output; otherwise, it returns to execute the step of selecting an APS rule for scheduling.
The simulation-based closed-loop APS scheduling optimization system according to claim 9, wherein the determining the weight configuration of the selected APS rule comprises:

A comprehensive particle swarm optimization algorithm is used to set the weight configuration for the selected APS rules; the comprehensive particle swarm optimization algorithm is improved on the basis of the basic particle swarm optimization algorithm, and it calculates the fitness value according to the key performance indicator data in the iterative optimization process Determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm, and use the simulated annealing algorithm to evaluate the optimal position of the particle itself, and use the mutation operation of the genetic algorithm to optimize the particle swarm. The position is evaluated and updated, the inertia weight value is dynamically adjusted, and finally the next generation particle population is generated by the update.
The simulation-based closed-loop APS scheduling optimization system according to claim 9, wherein the simulation-based scheduling module (30) comprises: a scheduling sub-module (31) and a simulation sub-module (32); wherein,

The scheduling submodule (31) is used to determine the APS rule selected for scheduling, and determine the weight configuration of the selected APS rule; generate a scheduling Gantt chart according to the weight configuration of the APS rule; derive the corresponding scheduling The order sequence of the Gantt chart is stored in the database; and the simulation result based on the order sequence is obtained from the database; it is judged whether the simulation result meets the needs, and if so, the corresponding scheduling Gantt chart is output; otherwise, return to execute The described operation of selecting APS rules for scheduling;

The simulation submodule (32) is used for loading the order sequence and preset simulation model configuration data from the database, and loading the order sequence and the simulation model configuration data into factory design simulation software In the basic simulation model of the module, a scheduling simulation model is obtained; the scheduling simulation model is run, and key performance indicator data is obtained by evaluation; and the key performance indicator data is stored in the database as a simulation result.
The simulation-based closed-loop APS scheduling optimization system according to claim 11, wherein the simulation-based scheduling module (30) further comprises: an optimization algorithm sub-module (33); The selected APS rule sets a weight configuration, and provides the weight configuration to the scheduling sub-module (31);

The comprehensive particle swarm optimization algorithm is improved on the basis of the basic particle swarm optimization algorithm. In the iterative optimization process, the fitness value is calculated according to the key performance indicator data to determine the optimal position experienced by the particle itself and the value experienced by the particle swarm. The optimal position, and the simulated annealing algorithm is used to evaluate the optimal position of the particle itself, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position of the particle swarm, dynamically adjust the inertia weight value, and finally update the next generation. particle population.
The simulation-based closed-loop APS scheduling optimization system according to claim 12, wherein the optimization algorithm sub-module (33) assigns initial weights to the selected APS rules, and performs particle encoding on the initial weights; The weights after particle encoding are initialized to obtain the current particle swarm; the current particle swarm is decoded to obtain the current weight configuration of the APS rule, and the current weight configuration is provided to the scheduling sub-module (31) Obtain the KPI data obtained by the evaluation of the simulation submodule (32), and calculate the fitness value according to the KPI data to determine the optimal position experienced by the particle itself and the optimal position experienced by the particle swarm; Determine whether the maximum number of iterations has been reached? When the maximum number of iterations is not reached, the simulated annealing algorithm is used to evaluate the optimal position experienced by the particle itself, and the mutation operation of the genetic algorithm is used to evaluate and update the optimal position experienced by the particle swarm; A new generation of the current particle swarm, and return to perform the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule; when the maximum number of iterations is reached, the optimal The particle swarm corresponding to the position is used as the current example swarm, and the operation of decoding the current particle swarm to obtain the current weight configuration of the APS rule is returned to execute.
The simulation-based closed-loop APS scheduling optimization system according to any one of claims 9 to 13, wherein the simulation-based scheduling module (30) further comprises: an APS rule configuration module (34) for Rules already existing in the advanced planning and scheduling software and/or custom rules added to determine the rules required for the current schedule.
A computer-readable storage medium on which a computer program is stored; characterized in that the computer program can be executed by a processor and implements simulation-based closed-loop advanced planning and scheduling software (APS) as described in claims 1 to 4 ) scheduling optimization method.