Application of Modified Benders Decomposition to Single-Stage Multi-Commodity Multi-Period Warehouse Location Problem: An Empirical Investigation  

作　　者：Renduchintala Raghavendra Kumar Sharma Ankita Malviya Vimal Kumar Vinay Singh Pritee Agarwal Renduchintala Raghavendra Kumar Sharma;Ankita Malviya;Vimal Kumar;Vinay Singh;Pritee Agarwal(Department of Industrial & Management Engineering, Indian Institute of Technology, Kanpur, India;ABV Indian Institute of Information Technology & Management, Gwalior, India;Department of Mathematics & Statistics, Indian Institute of Technology, Kanpur, India)

机构地区：[1]Department of Industrial & Management Engineering, Indian Institute of Technology, Kanpur, India [2]ABV Indian Institute of Information Technology & Management, Gwalior, India [3]Department of Mathematics & Statistics, Indian Institute of Technology, Kanpur, India

出　　处：《American Journal of Operations Research》2016年第3期245-259,共15页美国运筹学期刊（英文）

摘　　要：In the past, researchers have applied Bender’s decomposition to distribution problem and used feasibility constraint to speed up the performance of Bender’s decomposition. Further, the application of Branch and Bound to single-stage multi-commodity single-period warehouse location problem (SSMCSPWLP) with strong constraints has shown that they are more effective. It was also shown in the previous research (in the context of Branch and Bound Methodology) that hybrid formulation for the single-stage single-period multi-commodity warehouse location problem yielded superior results. In this paper we apply Benders’ decomposition to strong and weak formulations of single-stage multi-commodity multi-period warehouse location problem (SSMCMPWLP). As suggested in the previous literature we put feasibility constraints in the pure integer sub- problem to speed up the performance of Benders’ decomposition. We also develop an additional cut (constraint that is again added to pure integer sub-problem) and show that it further speeded up Benders’ Decomposition. This research led to the possibility of applying Benders’ Decomposition to the hybrid formulation of SSMCMPWLP in future.In the past, researchers have applied Bender’s decomposition to distribution problem and used feasibility constraint to speed up the performance of Bender’s decomposition. Further, the application of Branch and Bound to single-stage multi-commodity single-period warehouse location problem (SSMCSPWLP) with strong constraints has shown that they are more effective. It was also shown in the previous research (in the context of Branch and Bound Methodology) that hybrid formulation for the single-stage single-period multi-commodity warehouse location problem yielded superior results. In this paper we apply Benders’ decomposition to strong and weak formulations of single-stage multi-commodity multi-period warehouse location problem (SSMCMPWLP). As suggested in the previous literature we put feasibility constraints in the pure integer sub- problem to speed up the performance of Benders’ decomposition. We also develop an additional cut (constraint that is again added to pure integer sub-problem) and show that it further speeded up Benders’ Decomposition. This research led to the possibility of applying Benders’ Decomposition to the hybrid formulation of SSMCMPWLP in future.

关 键 词：Benders Decomposition SSMCMPWLP Strong and Weak Formulation Warehouse Location DISTRIBUTION 

分 类 号：O17[理学—数学]