الفهرس | Only 14 pages are availabe for public view |
Abstract In this thesis, a novel optimization approach is developed. This optimization algorithm is called ”Inner-Outer Array Genetic based algorithm with and without constraint relaxation technique. Genetic algorithm, a powerful random based method suffers from two drawbacks: pre-mature convergence and large number of iterations even with most common parameter settings. The developed algorithm has two capabilities: first, exploratory capability where two arrays (full versus fractional arrays) with two levels (for linear problems) or higher number of levels (for nonlinear problems) are used to embody the search space through choice of variable levels. The space is even explored further through Inner-Outer-Outer Array arrangement. The second phase, Exploitation, starts by reaching the local optimal solution in the parameter design stage. The second stage, Tolerance Design stage, zooms further around the local optimal solution through a tuning stage to reach the global optimal solution. Both pure Genetic Algorithm and Inner – Outer based Genetic algorithm are employed for a set of standard nonlinear problems with success. |