Markov approach for inventory control with meta-heuristics in intermittent demand environment

Demand variability directly affects inventory management. The variability of intermittent demand causes high lost sales or holding costs. While lost sales reduce customer satisfaction, keeping excessive stock also creates high costs for companies. This situation can be prevented with an appropriate inventory policy. In this study, a Markov-based proactive inventory management approach supported by metaheuristic methods is proposed in the inventory management of intermittent demands. The main contribution of the proposed approach is to find a lower and upper limit for stock by modeling the intermittent demands in the past period with the Markov process. With these optimized limits, it is aimed to balance the largest costs caused by intermittent demands, namely stock and lost sales costs. The intermittent demands used were randomly generated in 4 different sizes from small to large. The proposed approach contributes to inventory management by minimizing the negativities caused by demand variability through the Markov process. A mathematical model has been proposed for stock level optimization, but no feasible solution has been found. The mathematical model was transformed into a fitness function and a solution was provided with the Tabu Search Algorithm and Simulated Annealing. The inventory management process of intermittent demand was first evaluated without the Markov approach, and then the Markov approach was included in the process. The results showed that the Markov approach was a good tool for inventory management of intermittent demand. When the results were examined, the stock limits computed with the Markov process balanced the increased inventory cost and lost sales costs due to intermittent demand.