zbMATH — the first resource for mathematics

Improving the particle swarm optimizer by function “stretching”. (English) Zbl 1015.90064
Hadjisavvas, Nicolas (ed.) et al., Advances in convex analysis and global optimization. Honoring the memory of C. Caratheodory (1873-1950). Dordrecht: Kluwer Academic Publishers. Nonconvex Optim. Appl. 54, 445-457 (2001).
Summary: A new technique, named function “stretching”, for the alleviation of the local minima problem is proposed. The main feature of this technique is the usage of a two-stage transformation of the objective function to eliminate local minima, while preserving the global ones. Experiments indicate that combined with the particle wwarm optimizer method, the new algorithm is capable of escaping from local minima and effectively locate the global ones. Our experience is that the modified algorithm behaves predictably and reliably and the results were quite satisfactory. The function “stretching” technique provides stable convergence and thus a better probability of success to the method with which it is combined.
For the entire collection see [Zbl 0968.00020].

90C26 Nonconvex programming, global optimization
90C59 Approximation methods and heuristics in mathematical programming
90C56 Derivative-free methods and methods using generalized derivatives