"A Hybrid Genetic/Interval Algorithm for Semi-Infinite Optimization"

C. Guarino Lo Bianco, and A. Piazzi

IEEE Conf. on Decision and Contr., CDC96, Kobe, Japan, Dec. 11-13, 1996, Vol. 2, pp. 2136-2138.

This paper proposes a hybrid algorithm for solving constrained semi-infinite optimization problems. It is based on a partially elitistic genetic algorithm which uses an interval procedure to compute penalty terms in constructing the fitness function. Due to the deterministic nature of the interval procedure, which globally converges with certainty, a robust overall algorithm is obtained. Computational results are reported, including those relative to the optimal PID design for H2 minimax control of an uncertain plant.