An extensible evolutionary algorithm approach for inverse problems

Due to their variety, it is difficult to find an adequate technique which can solve a broad range of inverse problems. This paper presents an extensive evolutionary algorithm approach, which is suitable for solving inverse problems. The system developed by the authors is programmed on a object-oriented base, and various evolutionary algorithms, such as genetic algorithms and continuous evolutionary algorithms, can be easily implemented. Implementation of new algorithms can also be easily achieved. In addition, programming in JAVA language allows us to use it remotely. As numerical examples, optimum design of a micromachine and parameter identification of an inelastic constitutive model, which often fail with conventional techniques, are solved with the system and the result of the investigation show the applicability of the approach.