• Cancela Carollo, Alberto
• Dorado de la Calle, Julián
• Rabuñal Dopico, Juan Ramón
• Pazos Sierra, Alejandro

This work proposes a genetic algorithm (GA) based approach for the search of the Pareto optimal set of a multiobjective optimization problem. First the global population es dividided into various subpopulations. The algorithm operation consist of two phases: firstly each subpopulation tries to optimize a different objective; later the algorithm searches for good compromise solutions between objectives. Information is exchanged by means of the migration of individuals during the second phase. A weighted sum is used for fitness calculation. Weight vectors are randomly generated for each selection event, wich creates a wide range of search directions. The good behaviour of the proposed algorithm becomes visible in its application to some continuous problems.