Genetic algorithm approach for large scale quadratic programming of probabilistic supplier selection and inventory management problem

Abstract

Abstract This paper is focused on the effectiveness of existing metaheuristic optimization method, genetic algorithm (GA), in providing solutions to a large-scale integer quadratic programming of a probabilistic supplier selection and inventory management problem. The term “probabilistic” in this case refers to any problem that involves some uncertain parameters, approached by random variables (probabilistic parameter). We used the existing mathematical model of probabilistic supplier selection problem and inventory management provided in our previous works. This was done for optimization problem in a small-scale, which could be solved efficiently either by analytical or numerical method. We resolved this model with an extensive number of decision variable, indicated by the number of supplier and time period sufficient to use GA, conducted to analyse if the solution of decision variable, is reliable or not for application to the larger problem. We generated some random data to simulate the problem and the results, running over hundreds of computational experiments. The results showed that the decision obtained by the genetic algorithm was significantly different from the global optimal solution generated by the generalized reduced gradient (GRG) performed in LINGO 18.0. In conclusion, GA is not preferred in solving large-scale problems as regards supplier selection and inventory management.