A Meta-Heuristic Intercell Scheduling Approach Based On Genetic Programming Initial Population

The unique cellular manufacturing paradigm in the military equipment manufacturing industry of our country is discussed in this thesis.In this research area,intercell production,which is common,is essentially an intercell transfer problem,can not be avoided.Based on the analyses of the actual production and the related research status at home and abroad,a more proper intercell scheduling model considering limited transportation capacity is proposed.Firstly,the intercell scheduling model considering limited transportation capacity is described in detail and mathematically formulated.Secondly,based on the above model,a meta-heuristic intercell scheduling approach based on genetic programming initial population(MGPI)is proposed.MGPI has two stages,in the first stage,genetic programming(GP)is automatically evolved to get new heuristic rules based on the information of the machines and vehicles.The obtained heuristic rules are used to generate corresponding initial solutions.The initial group of MGPI consists of the GP-based initial solutions,human-designed-based initial solutions and random-based initial solutions.In the second stage,a genetic algorithm(GA)is developed to select appropriate solutions for the vehicles and machines.Compared with classical meta-heuristic algorithms,the proposed algorithm has higher optimization performance as well as computational efficiency,because it takes the properties of the model into consideration more properly.Finally,to verify the performance of the proposed algorithm,a parameter experiment and a series of comparison experiments are conducted.Experimental results show that the proposed algorithm has a significant advantage on optimization performance and computational efficiency over other common intercell scheduling algorithms,and is especially suitable for problems with large instance sizes.