An optimization framework for multistage production-distribution networks using genetic algorithms

Research in the optimal design of multi-echelon production-distribution networks has been focusing on two-echelon models, which comprise the location-allocation of plants and distribution centers subject to specific constraints. Research in two-echelon models could be for two-stage or three-stage optimization model. Currently, almost all the research is in the two-stage model. A three-stage model which integrates DC and plant location-allocation decisions with vendor allocation decisions is, however, a more accurate abstraction of the real world, since the prices and transportation costs of raw materials can vary significantly amongst the vendors, depending on their locations viz-a-viz the plants to be opened. This problem is NP-complete and consists of a large number of variables. This paper provides a solution methodology using genetic algorithm for the optimization of a three-stage, two-echelon, multi-product, capacitated single sourcing production-distribution model.