Network Optimization Approaches to Solve the Stochastic and Dynamic Facility Layout Problems and Reduce Supply Chain Costs
This thesis researches on the Dynamic Facility Layout Problem (DFLP) and the Stochastic and Dynamic Facility Layout Problem (SDFLP). The problems are extensions to the static or single-period facility layout problem (SFLP). They assume that there are fluctuations in the products’ demands and consequently in the flows of material (and/or final products) between facilities in a given planning horizon. Fluctuations in flows of material are also due to the introduction of new products, disasters, and other production and marketing changes impacting the supply chain. In the DFLP, the flows of material between facilities vary over time but they are assumed known. In the SDFLP, the flows between facilities are uncertain and may follow different random distributions. The objective of these problems is to find an assignment of facilities to locations at each period that optimizes the material handling cost and the facilities relocation cost. This thesis has three contributions. First, it assesses the accuracy and efficiency of a <i>Parallel Shortest Path (PSP)</i> algorithm developed by Kolla (2015) to solve the DFLP. Second, it tests the efficiency on formulating a <i>linear network model (LNM)</i> for the DFLP and solving it with the network simplex algorithm implemented in AMPL, a commercial mathematical programming language, through numerical experimentation. Third, this thesis proposes a constrained shortest path network model to solve the SDFLP and experiments with small size instances. The SDFLP network model is an extension of the DFLP model in Balakrishnan et al. (1992).
Balachandran, G. (2018). <i>Network optimization approaches to solve the stochastic and dynamic facility layout problems and reduce supply chain costs</i> (Unpublished thesis). Texas State University, San Marcos, Texas.