Stochastic inventory control in closed-loop hybrid systems with remanufacturing

Sang-Phil Kim, Purdue University

Abstract

As industrial attention on sustainability has grown increasingly, more manufacturing companies are adopting environmental practices such as recycling and remanufacturing. The production decisions become challenging in hybrid manufacturing/remanufacturing systems due to the complexity of closed-loop systems. In this dissertation, the periodic review inventory controls in the closed-loop hybrid manufacturing/remanufacturing systems are studied. We address three main problems as follows. i) Inventory control in hybrid system with fixed set-up cost. The multi-period inventory problems in a closed-loop hybrid manufacturing and remanufacturing system with fixed set-up costs are studied. Due to the fixed set-up costs, it is difficult to analyze the optimal structural properties. We consider two types of systems with a fixed set-up cost, one with joint set-up cost and another with switching set-up cost and provide an approximate policy structure for each type of set-up costs by combining make-up-to and (s, S) policies. The numerical results show that the approximate policies provide good near-optimal solutions in terms of expected profit. ii) Inventory control in hybrid system with stock-out based substitution. Remanufacturing and manufacturing systems for a certain type of products can be considered independently when they have distinct markets. However, an interconnection between them could arise by product substitutions. Though such substitutions have benefits such as reductions in product shortage and inventory size, the discounted price may incur some losses in profit. We find that the optimal policy for production is the state dependent make-up-to policy where the optimal make-up-to level depends on the current inventory levels. iii) Joint promotion/inventory decision in hybrid system with expediting. The joint promotion/inventory decision problem in a closed-loop hybrid system is modeled as a stochastic dynamic programming. The demands are assumed to be stochastic and influenced by the promotion decision. The stochastic returns are also stochastic and correlated with the volume of circulation. The more product sold in the current period, the more returns in the future periods which leads to cost savings. For a given decision on the promotion, we conjecture the optimal solution structure and monotonicities of control parameters.

Degree

Ph.D.

Advisors

Lee, Purdue University.

Subject Area

Industrial engineering|Operations research

COinS