TY - JOUR
T1 - Multi-objective simulation optimization using data envelopment analysis and genetic algorithm
T2 - Specific application to determining optimal resource levels in surgical services
AU - Lin, Rung Chuan
AU - Sir, Mustafa Y.
AU - Pasupathy, Kalyan S.
N1 - Funding Information:
This work was supported in part by a generous grant from the University of Missouri Research Board.
PY - 2013/10
Y1 - 2013/10
N2 - Simulation is a powerful tool for modeling complex systems with intricate relationships between various entities and resources. Simulation optimization refers to methods that search the design space (i.e., the set of all feasible system configurations) to find a system configuration (also called a design point) that gives the best performance. Since simulation is often time consuming, sampling as few design points from the design space as possible is desired. However, in the case of multiple objectives, traditional simulation optimization methods are ineffective to uncover the efficient frontier. We propose a framework for multi-objective simulation optimization that combines the power of genetic algorithm (GA), which can effectively search very large design spaces, with data envelopment analysis (DEA) used to evaluate the simulation results and guide the search process. In our framework, we use a design point's relative efficiency score from DEA as its fitness value in the selection operation of GA. We apply our algorithm to determine optimal resource levels in surgical services. Our numerical experiments show that our algorithm effectively furthers the frontier and identifies efficient design points.
AB - Simulation is a powerful tool for modeling complex systems with intricate relationships between various entities and resources. Simulation optimization refers to methods that search the design space (i.e., the set of all feasible system configurations) to find a system configuration (also called a design point) that gives the best performance. Since simulation is often time consuming, sampling as few design points from the design space as possible is desired. However, in the case of multiple objectives, traditional simulation optimization methods are ineffective to uncover the efficient frontier. We propose a framework for multi-objective simulation optimization that combines the power of genetic algorithm (GA), which can effectively search very large design spaces, with data envelopment analysis (DEA) used to evaluate the simulation results and guide the search process. In our framework, we use a design point's relative efficiency score from DEA as its fitness value in the selection operation of GA. We apply our algorithm to determine optimal resource levels in surgical services. Our numerical experiments show that our algorithm effectively furthers the frontier and identifies efficient design points.
KW - DEA
KW - Genetic algorithm
KW - Multicriteria simulation optimization
KW - Surgical services
UR - http://www.scopus.com/inward/record.url?scp=84871765496&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84871765496&partnerID=8YFLogxK
U2 - 10.1016/j.omega.2012.11.003
DO - 10.1016/j.omega.2012.11.003
M3 - Article
AN - SCOPUS:84871765496
SN - 0305-0483
VL - 41
SP - 881
EP - 892
JO - Omega (United Kingdom)
JF - Omega (United Kingdom)
IS - 5
ER -