Interval type-2 fuzzy inference system for tuning adaptive weighted multi-objective genetic algorithm

Abstract

Many real world optimization problems involve multi-objectives. Multi-objective problems are problem with two or more objectives and generally conflicting with each other. Multi-objective optimization algorithms goals are converge to the set of Pareto optimal solutions and maintain of diversity among Pareto optimal solutions. Multi-objective optimization approaches can be divided into classical approaches and evolutionary algorithms. Classical approaches generally convert multi-objective function into single objective function and involve decision makers in the search. Evolutionary optimization algorithms use a population based approach in which a set of solutions evolves new solutions in the next generation. The use of population of solutions helps to simultaneously find a set of Pareto optimal solution, thus making evolutionary optimization computationally efficient. Genetic algorithm parameter is the key factor to determine genetic algorithm performance.