Improved Discrete Spider Monkey Optimization Using a Dynamic Penalty Function with Partial Constraint Violation Acceptance for Vending Machine Column Optimization Considering Sales and a Replenishment Cycle

This paper proposes an improved discrete spider monkey optimization based method using a dynamic penalty function with partial constraint violation acceptance for vending machine column optimization considering sales and a replenishment cycle. The conventional method has a challenge that it only searches for feasible regions. To address the challenge, the proposed method added a dynamic penalty function in order to search for infeasible regions. Moreover, partial constraint violation acceptance rate is introduced to obtain high quality solutions. Effectiveness of the proposed method is verified by comparing with the conventional method, a method with only a static penalty function, a method with only a dynamic penalty function with actual vending machine data.