Global Optimization Based on Weighting-Integral Expected Improvement
Keyword(s):
Balancing the global exploration and the local exploitation has received particular attention in global optimization algorithm. In this paper, based on Kriging model an infill sample criteria named weighting-integral expected improvement is proposed, which provides a high flexibility to balance the scope of search. Coupled with this infill sample criteria, a strategy is proposed that on each iteration step, the infill sample point was selected by the urgency of each search scope. Two mathematical functions and one engineering problem are used to test this method. The numerical experiments show that this method has excellent efficiency in finding global optimum solutions.
2015 ◽
Vol 2015
◽
pp. 1-15
◽
Keyword(s):
2018 ◽
Vol 108
(07-08)
◽
pp. 499-505
2008 ◽
Vol 2
(1)
◽
pp. 49-55
◽