Application of data mining in a global optimization algorithm

Abstract

A hybrid global optimization algorithm is developed in this research. The probability of finding the global optimal solution is increased by reducing the search space. The activities of classification, association, and clustering in data mining are employed to achieve this purpose. The hybrid algorithm developed uses data mining (DM), evolution strategy (ES) and sequential quadratic programming (SQP) to search for the global optimal solution. For unconstrained optimization problems, data mining techniques are used to determine a smaller search region that contains the global solution. For constrained optimization problems, the data mining techniques are used to find the approximate feasible region or the feasible region with better objective values. Numerical examples demonstrate that this hybrid algorithm can effectively find the global optimal solutions for two benchmark test problems.