Robust Optimization Approaches for Assemble-To-Order Systems

Abstract

This research studies an Assemble-To-Order (ATO) system affected by demand and lead time uncertainty. Firstly, we use the classical Robust Optimization (RO) approach to protect the system against perturbation and solve the component allocation problem to optimality. Secondly, we design a robust two-stage model using the Adaptive Robust Optimization (ARO) methodology to solve the joint optimization problem of base-stock level and component allocation. This multi-stage problem is solved using the Column and Constraint Generation (CCG) method upon discovering specific characteristics of the problem. Our study incorporates the concept of the budget of uncertainty, determined by the decision maker, allowing them to be as flexible as they desire in accounting for uncertainty in the system. Lastly, the correlation between the total reward and the base stock levels and overall budget is investigated for different values of budgets of uncertainty, and numerical results are reported.