Genetic Algorithm For Supply Planning Issues

Supply planning optimization is one of the most important issues for manufactures and distributors. Supply is planned to meet the future demand. Under the uncertainty involved in demand forecasting,profit is maximized and risk is minimized.TEZUKA Masaru and HIJI masahiro gave an genetic algorithm to solve the supply planning problems Under the uncertainty,independence involved in demand forecasting.However,the demand of products is correlated in fact. So in this paper the author repaired it in order to satisfy the correlation situation. When using Monte carlo simulation to evaluate the profit and risk,the author assumed every product following Normal Distribution,and there were correlation random number between them which following Normal Distribution.Finally the random number were used as the future demand.This paper also did the statistical analysis for the algorithm by sampling.and obtain a good result.In the process of generating initial population ,the author introduced Michalewicz Boundary handling technique ,and repaired it in order to resolve the constraint condition of supply planning problem .When Comparing two individuals ,the author proposed a policy which compared vector fitness.And this paper proposed how to prevent early-maturing mechanism. The approach was tested on the supply planning data of an electric appliances manufacturer,and has achieved a remarkable result.