Genetic Algorithm Parameters in a Vendor Managed Inventory Model

To handle the information exchange problem between a vendor and a retailer, Vendor Managed Inventory (VMI) provides a good approach to handle the problem. Information exchanges between both sides enhance supply chain performance. In a previous research work, a stochastic model for one vendor and one retailer has been developed. Simulation optimization using genetic algorithm (GA) has been use to solve the problem. There are 2 important parameters in genetic algorithm (probability of mutation and probability of crossover). This research aims at analyzing relations between GA parameters and optimal solutions. This research compares many combinations of GA parameters and the effects on optimal solutions and time to reach the optimal solutions. This research concludes that the best combination reaches the optimal solutions. Unfortunately, the best combination is only suitable for a certain condition and increasing/reducing GA parameters values do not automatically improve the optimal solutions.