| With the development of the intelligent era,traditional manufacturing enterprises are gradually transforming into smart manufacturing enterprises.Meanwhile,smart manufacturing systems such as manufacturing execution system(MES)used by enterprises need to take into account their logistics,production and other links,which directly affect their benefits.Therefore,it is crucial for MES systems to obtain efficient scheduling schemes and efficient algorithms that include all links.In addition,the production mode is gradually changing to distributed production,and there are often complex loading and distribution processes among the job produced in distributed factories.In addition,there are many uncertainties in the transportation and production processes of actual production activities.It is more practical and instructive to obtain efficient scheduling schemes in uncertain environments.Therefore,it is of great economic value and theoretical interest to study the vehicle distribution considering load and distributed production integrated scheduling problem.In this paper,we focus on modeling and solving the vehicle distribution considering load and distributed production integrated scheduling problem.The main points are as follows:(1)On the basis of the traditional distributed permutation flow-shop scheduling problem(DPFSP),further considering the process that the workpiece processed in distributed factory is loaded in the raw material warehouse and distributed to the factory by vehicle,the capacitated vehicle distribution and distributed production integrated scheduling problem(CVDDPISP)model is established for the first time to minimize the cost of distribution and production stage and as an optimization objective.Meanwhile,an improved three-dimensional estimation of distribution algorithm(I3DEDA)is proposed for solving the problem.First,based on the characteristics of the CVDDPISP problem,a codec rule is designed by using a cost balancing strategy at each stage.Secondly,in order to improve the global search ability of the algorithm,the block structure and location information of high-quality encoded individuals in the population were learned and accumulated by using three-dimensional estimation of distribution algorithm(3DEDA),and then the probabilistic model was sampled to generate new population individuals.In addition,an effect local search is designed to improve the performance of the three-dimensional distribution estimation algorithm.Finally,simulation experiments and algorithms are compared to verify that I3 DEDA can solve CVDDPISP efficiently.(2)On the basis of(1),considering the complex loading process between fragile and fragile workpieces in actual production activities,the vehicle distribution with twodimensional loading constraints and distributed production integrated scheduling problem was established for the first time with the optimization goal of minimizing the cost of loading,distribution,and production at three stages.At the same time,a hybrid3DEDA(H3DEDA)is proposed for solving the problem.Based on(1),this algorithm designs a two-dimensional loading strategy which is beneficial for decision-making optimization of each stage according to the job characteristics.Combined with the structural characteristics of VD2 LDPISP loading,distribution and production,the codec rules of(1)are improved.In addition,in order to enhance the local optimization ability of the algorithm,a high-low hierarchical super-heuristic local search hyperheuristic local search(HHLS)is designed,which uses a two-dimensional probability model to reasonably learn high-quality neighborhood operation order information and then dynamically guide the search direction of H3 DEDA,which can avoid the problem of insufficient search depth caused by unreasonable neighborhood structure to a certain extent.Finally,we validate the effectiveness of H3 DEDA through simulation experiments with different scale examples.(3)On the basis of(2),fuzzy vehicle distribution with two-dimensional loading constraints and distributed production integrated scheduling problem(FVD2LDPISP)was established for the first time by further considering the uncertainties caused by various factors affecting and restricting the actual production activities,causing the vehicle speed on the road section and the job processing time on the machine.In addition,the triangular fuzzy number is used to deal with the uncertain value,and the fuzzy number calculation correction scheme is used to correct the fuzzy number that violates the constraint.Meanwhile,an enhanced 3DEDA(E3DEDA)was designed to solve the problem.In order to take into account the global search ability and local search efficiency of the algorithm,the algorithm first designed a decoding strategy based on the cost balance of each stage according to the characteristics of FVD2 LDPISP to decode the encoding individual containing the information of the job and the production factory.Second,in order to reduce the complexity of the three-dimensional probabilistic model,the structure and location information of the blocks of jobs and production plants in the high-quality encoded individuals is accumulated using the three-dimensional probabilistic model to reasonably guide the evolution of the population.Then,on the basis of HHLS in(2),combined with the record of algorithm coding and 3D probability model information,a variety of neighborhood operations and update evaluation methods based on three-dimensional probability model are designed to improve the efficiency of algorithm search.Finally,experiments and algorithms are performed for problems of different sizes to verify that E3 DEDA is an effective algorithm for solving FVD2 LDPISP. |
