Supply Chain Network Design with Response Time

Consideration: Models and Solution Methods

 

Navneet Vidyarthi, Samir Elhedhli, & Elizabeth Jewkes

Department of Management Sciences

University of Waterloo

200 University Avenue West

Waterloo, ON, N2L 3G1, Canada.

{nkvidyar, elhedhli, emjewkes}@uwaterloo.ca

 

Make-to-order and assemble-to-order systems are important business models in managing a wide ranging class of responsive supply chains, characterized by highly uncertain customer demand, high product variety, and short product life cycle. Motivated by the strategic importance of response time/lead time reduction in supply chain management, the primary focus of the proposed research is to develop models for designing make-to-order and assemble-to-order supply chains. We consider a two-echelon make-to-order supply chain design model characterized by stochastic customer demand that has to be satisfied from a set of distribution centers (DCs), where sufficient production/assembly capacity has to be acquired in order to avoid long response time. The objective is to simultaneously determine the DC locations, capacity levels at DCs, and the assignment of customers to DCs by minimizing response time costs in addition to the variable transportation cost, fixed cost of opening DCs, and capacity acquisition costs at DCs.  The DCs are modelled as spatially distributed M/G/1 queue. The model is formulated as a non-linear mixed integer programming (MIP) problem. The model exhibits nonlinearity because of the integration of response time in optimizing network design. We propose a linearization based on a simple transformation and piece-wise linear and concave approximations. We present an exact solution method based on cutting plane approach which is amenable to state-of-the-art IP solvers. Preliminary computational results demonstrate that the cutting plane approach provides the optimal solution in few iterations

 

Back to Presenters’ Page