Multi-objective optimisation with stochastic algorithms and fuzzy definition of objective function

Article type: Research Article

Authors: Chiampi, M.^a | Fuerntratt, G.^b | Magele, Ch.^{b; *} | Ragusa, C.^a | Repetto, M.^a

Affiliations: [a] Dipartimento di Ingegneria Elettrica Industriale, Politecnico di Torino, Torino, Italy | [b] IGTE, Technische Univeristät Graz, Graz, Austria

Correspondence: [*] Corresponding author. Fax: +43 316 873 7751; E-mail: magele@igte.tu@graz.ac.at.

Abstract: The use of a decision making scheme based on fuzzy logic applied to the design of electromagnetic devices is presented. In the Multi-Objective Optimisation, the different scalar objectives are merged together firstly by means of a fuzzyfication process and then by the application of a fuzzy logical operator. This scheme can be easily coupled to any search optimisation (zeroth order algorithm) since it does not require differentiation of the scalar functions. Two different implementations of the procedure, coupled to Simulated Annealing and to Genetic Algorithm stochastic search are presented. The two procedures are tested versus a two solenoids SMES (Superconducting Magnetic Energy Storage) Multi-Objective Optimisation problem.

DOI: 10.3233/JAEM-1998-097

Journal: International Journal of Applied Electromagnetics and Mechanics, vol. 9, no. 4, pp. 381-389, 1998