A System-Dynamic Based Simulation and Bayesian Optimization for Inventory Management

Abstract

Inventory management is a fundamental challenge in supply chain management. The challenge is compounded when the associated products have unpredictable demands. This study proposes an innovative optimization approach combining system-dynamic Monte-Carlo simulation and Bayesian optimization. The proposed algorithm is tested with a real-life, unpredictable demand dataset to find the optimal stock to meet the business objective. The findings show a considerable improvement in inventory policy. This information is helpful for supply chain analytics decision-making, which increases productivity and profitability. This study further adds sensitivity analysis, considering the variation in demand and expected output in profit percentage. This paper makes a substantial contribution by presenting a simple yet robust approach to addressing the fundamental difficulty of inventory management in a dynamic business environment.