The Research On Production Scheduling Problems Of The Mixed-model Assembly Line Based On Genetic Algorithms

In order to respond quickly to the market and meet the growing diversified andpersonalized demand of customers, more and more manufacturing enterprises adoptthe mixed-model assembly production mode to realize flexible manufacturing. Inthis manufacturing environment, formulating a reasonable production sequencebecomes one of the important point through which enterprises reduce theirmanufacturing cost.When doing researches on the production scheduling problem，this paperrespectively builds the single-objective optimization model of minimizing thebalanced consumption degree of the parts and components and minimizing thecompletion time. the purpose is to lower the inventory levels of the mixed-modelassembly production mode at the corresponding production stratum, realize levelproduction, and increase the production capacity and the efficiency of enterpriseswithout inputs of more production factors.To balance the consumption of the partsand components, the paper controls the degree of the balance level of various partsby introducing weight, this method gives priority to parts and components whichhave a great impact on the production system when they are not balanced. At thesame time, this article will consider both the adjustment time and the assemblyoperation time to build the completion time model at a loop period, trough thismethod, a more accurately total completion time can be obtained. Then this papergives a comprehensive consideration to the above two single objectives ofproduction scheduling to construct the multi-objective optimization model and usesthe weight coefficient method to transform the multi-objective optimization problemto the single-objective optimization problem.This paper uses the genetic algorithm to solve the transformedsingle-objective optimization problem. On the basis of the standard genetic algorithm, this article builds the initial population by selecting better individualsfrom several population which is randomly generated in proportion, this buildingmethod of the initial population inspired from the niche technology can improve thediversity of the initial population and the optimization speed of the geneticalgorithm.When selecting the replication operator, the paper combines the rouletteand elite strategies together to reserve the optimal individual and enhance theconvergence performance.Finally, by using the calculation function of the MATLAB software, theconstructed production scheduling model and the genetic algorithm model，the paperobtains the vehicle production sequence in a particular day for F company’s vehiclemixed-model assembly line, and through the comparison and analysis between theproduction sequence calculated by this paper and the production sequence currentlyusing in F company, it shows that the product sequence this paper obtains can lowerthe balanced consumption degree of the parts and components obviously and shortenthe total completion time required for assembling the same quantity of vehiclesapparently.