TY - JOUR
T1 - Reduced search space mechanism for solving constrained optimization problems
AU - Sallam, Karam M.
AU - Sarker, Ruhul A.
AU - Essam, Daryl L.
N1 - Publisher Copyright:
© 2017 Elsevier Ltd
PY - 2017/10/1
Y1 - 2017/10/1
N2 - Over the last few decades, a considerable number of evolutionary algorithms (EAs) have been proposed for solving constrained optimization problems (COPs). As for most of these problems, the optimal solution exists on the boundary of the feasible space, we aim to focus the search process around the boundary. In this paper a new concept, called reduced search space (R2S), is introduced. In the process, we first identify active constraints, based on the current solutions, and then define R2S around those constraint's boundaries. However, the search may be conducted either in the entire R2S or in some portions of it. To judge the impact of this concept, we have incorporated it with a number of state-of-the-art algorithms, and we have comprehensively tested it on three sets of benchmark test functions, namely, 24 test functions taken from IEEE CEC2006, 18 test functions with 10D and 30D taken from IEEE CEC2010 and 10 test functions taken from IEEE CEC2011. The results show that our proposed mechanism significantly improves the performances of state-of-the-art algorithms.
AB - Over the last few decades, a considerable number of evolutionary algorithms (EAs) have been proposed for solving constrained optimization problems (COPs). As for most of these problems, the optimal solution exists on the boundary of the feasible space, we aim to focus the search process around the boundary. In this paper a new concept, called reduced search space (R2S), is introduced. In the process, we first identify active constraints, based on the current solutions, and then define R2S around those constraint's boundaries. However, the search may be conducted either in the entire R2S or in some portions of it. To judge the impact of this concept, we have incorporated it with a number of state-of-the-art algorithms, and we have comprehensively tested it on three sets of benchmark test functions, namely, 24 test functions taken from IEEE CEC2006, 18 test functions with 10D and 30D taken from IEEE CEC2010 and 10 test functions taken from IEEE CEC2011. The results show that our proposed mechanism significantly improves the performances of state-of-the-art algorithms.
KW - Boundary search
KW - Constrained optimization problem
KW - Differential evolution
KW - Evolutionary algorithms
KW - Reduced search space
UR - http://www.scopus.com/inward/record.url?scp=85027588161&partnerID=8YFLogxK
U2 - 10.1016/j.engappai.2017.07.018
DO - 10.1016/j.engappai.2017.07.018
M3 - Article
SN - 0952-1976
VL - 65
SP - 147
EP - 158
JO - Engineering Applications of Artificial Intelligence
JF - Engineering Applications of Artificial Intelligence
ER -