Track: Facilities Planning and Plant Layout

Abstract

Plant location is a crucial decision in meeting the customer demand efficiently and effectively. The researchers have given different formulations for solving this optimization problem and such problems are NP Hard due to various integer variables and non linear formulation. Sharma and Muralidhar (2009) approach to model relaxed Simple Plant Location Problem (SPLP) resulted in two weak formulations and a strong formulation. It is found that weak formulations give an inferior bound compared to strong formulation but their computational time is comparatively very less. The reason being the constraints in strong formulations go quadratic in number. It has been observed by many that for smaller sized problems, the weak formulation of SPLP is superior to Strong formulation of SPLP in terms of computational time. Here hybrid formulations are proposed to accommodate benefits of both the formulations. We take full weak formulation (either weak one or weak two) and add to it only the most promising strong constraints. It is claimed that this approach (hybrid) is better than weak and strong formulation of SPLP in terms of upper bound and computational time. We also provide additional strong constraints based on various transportation and location costs assumptions that can improve the relaxed solutions of the SPLP.