zbMATH — the first resource for mathematics

Generalized simulated annealing for function optimization. (English) Zbl 0609.65045
The authors describe a heuristic approach to find the global minimizer of an arbitrary real-valued function, where constraints may be included. A search direction is determined randomly and a new iterate is accepted, if the function value is improved or if at least a test of the form $$V<\exp (-\beta f(x_ k)^ g(f(x_{k+1})-f(x_ k)))$$ is satisfied, V randomly chosen. Numerical results obtained by an academic example indicate that the algorithm is quite sensitive to the choice of parameters. A more detailed investigation of a practical optimization problem (concentration of a neurotransmitter) and a comparison with existing results are presented.
Reviewer: K.Schittkowski

MSC:
 65K05 Numerical mathematical programming methods 90C30 Nonlinear programming
Full Text: