Intelligent Optimization Algorithms For Unrelated Parallel Machine Scheduling Problem

In the context of industrial Internet,China’s manufacturing system is constantly being upgraded and moving towards informationization,digitization and intelligence,striving to build an integrated intelligent collaborative manufacturing mode of products,resources,manufacturing and services.With the diversified customized demands,changing production environment and complex process constraints,the flexible production and manufacturing mode with multiple varieties and small batches is becoming increasingly prevalent,which poses certain challenges to the establishment of the scheduling system.As a critical component of the manufacturing system,scheduling is an essential way to enhance the flexibility of the scheduling system in accordance with the intelligent collaborative manufacturing mode.In view of the actual production situation of enterprises,this paper focuses on the research of uncorrelated parallel machine scheduling,proposing a multi-constrained uncorrelated parallel machine scheduling problem with various availability constraints,particularly taking into account the product lot constraints,and solving it based on simulated annealing algorithm and group teaching optimization algorithm in combination with problem constraint design strategy,and finally applying the algorithm to the actual problem solving of Company J.The primary research contents of this paper are as follows:(1)For the uncorrelated parallel machine scheduling problem incorporating setup time,the simulated annealing algorithm based on the critical path-based neighborhood search approach(SA-CP)is employed for solving the problem.To this end,a mixed integer programming model with the goal of minimizing the maximum completion time is initially formulated through problem analysis,and a coding method relying on the composite of real numbers and weights is adopted due to the problem characteristics.Moreover,a heuristic rule for the critical path-based neighborhood search approach is proposed,which is then incorporated into the search operator of the simulated annealing algorithm to enhance the search efficiency.Lastly,simulations are conducted to compare SA-CP with other algorithms for tackling unrelated parallel machine scheduling problems involving setup time,thereby verifying the efficiency of SA-CP in dealing with such problems.(2)For the uncorrelated parallel machine scheduling problem,which considers product lot constraints and machine suitability constraints,a problem scheduling model is proposed with the optimization objective of minimizing the maximum completion time.This model combines lot scheduling with uncorrelated parallel machine scheduling and takes into account machine suitability constraints.To address the issue of the group teaching optimization algorithm converging too fast,a hybrid group teaching optimization algorithm(HGTOA)is proposed that incorporates the Metropolis criterion of the simulated annealing algorithm.Weight-based coding is used to continuousize the discrete scheduling problem,and the incorporation of the Metropolis criterion ensures the effectiveness of the HGTOA.The simulation experiments confirm the effectiveness of the HGTOA.(3)Taking a domestic textile manufacturing company’s weaving sheet workshop as an example,we summarize the scheduling requirements,integrate the setup time constraint,machine suitability constraint and product lot constraint to construct a scheduling model for the unrelated parallel machine problem,which is subject to multiple availability constraints.Subsequently,a hybrid grouping teaching optimization algorithm based on the double weight encoding method is proposed to solve the problem,and its effectiveness is confirmed through simulation experiments.Finally,the proposed HGTOA is applied to solve practical production cases.In this paper,two algorithms,namely SA-CP and HGTOA,are proposed for addressing the uncorrelated parallel machine scheduling problem with various constraints.The performance of the algorithms is validated by executing simulation experiments.Furthermore,the effectiveness and practicability of the algorithms are proven by solving the practical production cases of J Company,which integrates multiple constraints.