Approaches For Solving M-Best 3-Dimensional Dynamic Scheduling Problems For Large M
2016 19TH INTERNATIONAL CONFERENCE ON INFORMATION FUSION (FUSION)(2016)
摘要
Dynamic scheduling problems arise in areas, including traveling salesman problems, vehicle routing, nuclear fuel assembly loading patterns, and asset scheduling for resource allocation problems, to name a few. A generalized 3-dimensional scheduling problem is a decision making process that involves allocating limited resources to a set of tasks over time, where the objective is to optimize a cost function subject to realistic constraints. In addition, the realistic constraints complicate the problem further due to a possibly large number of disconnected feasible solution subspaces. These 3-dimensional assignment problems are known to be NP-hard. In this paper, we investigate a 3-dimensional assignment problem widely applicable to many fields (e.g., teacher scheduling, asset routing/scheduling, nuclear fuel management, multi-target tracking). We propose a novel approach to efficiently solve an m-best 3-dimensional scheduling problem, where m may be large (up to 1 0 4), by decomposing it into two sequential phases. In phase I, we partition the original problem space into a series of subproblems via Murty's m-best search space decomposition procedure. In phase II, we solve each subproblem using a combination of relaxed 2-dimensional assignments through reformulation into either a simplex-based transportation or transauction problem. We follow this two-step approach and generate large numbers of solutions in a relatively short amount of time. We validate and compare the performance of our proposed algorithm via experimental results.
更多查看译文
关键词
m-best 3D dynamic scheduling problems,traveling salesman problems,vehicle routing,nuclear fuel assembly loading patterns,asset scheduling,resource allocation problems,generalized 3D scheduling problem,decision making process,3D assignment problems,NP-hard,Murty m-best search space decomposition procedure,relaxed 2D assignments,simplex-based transportation,transauction problem
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要