Global optimization with a limited solution time

In this paper the box constrained global optimization problem in presence of a limited solution time is considered. A method is studied based on a combination of multistart and singlestart which implies a decision sequence on the number of random points to be generated. Search strategies are numerically illustrated. Criteria are introduced to measure the performance of solution methods for the problem class. Moreover, the performance of search strategies, specifically the efficiency of generating random points is analyzed.