A Partition-Based Random Search for Stochastic Constrained Optimization via Simulation

被引:18
作者
Gao, Siyang [1 ]
Chen, Weiwei [2 ]
机构
[1] City Univ Hong Kong, Dept Syst Engn & Engn, Kowloon, Hong Kong, Peoples R China
[2] Rutgers State Univ, Dept Supply Chain Management, Newark, NJ 07102 USA
关键词
Feasibility detection; partition-based random search (PRS); simulation budget allocation; simulation optimization; stochastic constraints; DISCRETE OPTIMIZATION; NESTED PARTITIONS; ALLOCATION;
D O I
10.1109/TAC.2016.2570119
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We consider the global optimization problem over finite solution space with a deterministic objective function and stochastic constraints, where noise-corrupted observations of the constraint measures are evaluated via simulation. This problem is challenging in that the solution space often lacks rich structure that can be utilized in identifying the optimal solution, and the feasibility of a solution cannot be known for certain, due to the noisy measurements of the constraints. To tackle these two issues, we adopt a partitioning scheme to explore the solution space and develop a feasibility detection procedure to detect the feasibility of the sampled solutions. A new random search method, called partition-based random search with multi-constraint feasibility detection (PRS-MFD), is proposed. It is shown that PRS-MFD converges to the set of global optima with probability one. The significantly higher efficiency of it is demonstrated by numerical experiments.
引用
收藏
页码:740 / 752
页数:13
相关论文
共 40 条
[1]   Optimizing discrete stochastic systems using simulated annealing and simulation [J].
Ahmed, MA ;
Alkhamis, TM ;
Hasan, M .
COMPUTERS & INDUSTRIAL ENGINEERING, 1997, 32 (04) :823-836
[2]   Simulation optimization for an emergency department healthcare unit in Kuwait [J].
Ahmed, Mohamed A. ;
Alkhamis, Talal M. .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2009, 198 (03) :936-942
[3]   A simulated annealing algorithm with constant temperature for discrete stochastic optimization [J].
Alrefaei, MH ;
Andradóttir, S .
MANAGEMENT SCIENCE, 1999, 45 (05) :748-764
[4]   Discrete stochastic optimization using variants of the stochastic ruler method [J].
Alrefaei, MH ;
Andradóttir, S .
NAVAL RESEARCH LOGISTICS, 2005, 52 (04) :344-360
[5]   A modification of the stochastic ruler method for discrete stochastic optimization [J].
Alrefaei, MH ;
Andradóttir, S .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2001, 133 (01) :160-182
[6]   Finding the best in the presence of a stochastic constraint [J].
Andradóttir, S ;
Goldsman, D ;
Kim, SH .
PROCEEDINGS OF THE 2005 WINTER SIMULATION CONFERENCE, VOLS 1-4, 2005, :732-738
[7]  
Andradóttir S, 2006, HBK OPERAT RES MANAG, V13, P617, DOI 10.1016/S0927-0507(06)13020-0
[8]   Balanced Explorative and Exploitative Search with Estimation for Simulation Optimization [J].
Andradottir, Sigrun ;
Prudius, Andrei A. .
INFORMS JOURNAL ON COMPUTING, 2009, 21 (02) :193-208
[9]   Finding Feasible Systems in the Presence of Constraints on Multiple Performance Measures [J].
Batur, Demet ;
Kim, Seong-Hee .
ACM TRANSACTIONS ON MODELING AND COMPUTER SIMULATION, 2010, 20 (03)
[10]   Selecting a selection procedure [J].
Branke, Juergen ;
Chick, Stephen E. ;
Schmidt, Christian .
MANAGEMENT SCIENCE, 2007, 53 (12) :1916-1932