Research On Two-stage Production Scheduling Optimization Based On ALNS

With the improvement of residents’ living standards and health awareness,China’s liquid daily necessities industry market has maintained a stable development trend,and the market scale has expanded year by year,but the manufacturing process of liquid daily necessities has become increasingly complex,and the problems of waste of production capacity and imbalance between supply and demand are particularly prominent.In addition,there are problems such as order prioritization,resource constraints and various unexpected situations in manufacturing.With the demand of digital transformation,the importance of production scheduling is particularly prominent,reasonable production scheduling can reduce the production cost of enterprises,improve the production efficiency of enterprises,and the research of related issues is particularly important for liquid daily necessities enterprises.This paper takes Company D as the research object and investigates the actual production environment of liquid daily necessities of Company D.The original production schedule of enterprise D is realized manually,which has problems such as dealing with constraint bottlenecks,consuming a lot of time and human resources,and difficulty in meeting the needs of high sudden changes.The production process of liquid daily necessities mainly includes liquid firing and liquid filling,and there are multiple parallel machines of different qualities in each process.The problems faced by the production of liquid daily necessities in enterprises can be classified as Hybrid Flow Shop Scheduling Problem,which is a typical NP-hard problem.Based on the investigation and analysis of the production scheduling problem of enterprise D,this paper constructs a mixed integer programming model of the two-stage production schedule of enterprise D based on the actual production situation of enterprise D,combined with the buffer constraint of transit area and production scheduling,and aims to minimize the maximum completion time.Therefore,a super-heuristic algorithm based on Adaptive Large Neighborhood Search(ALNS)is designed to solve and optimize the model.The algorithm adopts the ALNS algorithm in high-level strategies;Combined with the characteristics of two-stage production scheduling of D enterprises,six low-level heuristic methods are designed.The algorithm first adopts the heuristic method to construct the initial solution,and performs pu iterations based on the weights of low-level heuristic operators,then uses the threshold acceptance algorithm as the acceptance criterion,scores the current solution,and finally performs the next round of iterations according to the new weights until the maximum number of iterations is reached.In addition,combined with the scheduling needs of enterprise D in the second stage,this paper further designs two algorithm strategies to optimize the comprehensive application effect of the algorithm in the two stages.It has been verified that the ALNS super-heuristic algorithm has a significant improvement compared with the simulated annealing algorithm,tabu search algorithm and the manual scheduling method of D enterprise,which not only has a faster convergence speed in medium and large-scale cases,is easier to jump out of the local optimization of the algorithm,but also has a short production time in the magnitude of large-scale examples.It effectively alleviates the contradiction of enterprise D in the production scheduling problem,shortens the production cycle,optimizes the equipment operation rate,and effectively improves the overall production efficiency and comprehensive competitiveness of the enterprise.