A distance-based pareto evolutionary algorithm based on SPEA for combinatorial problems

In 1999, the so-called Strength Pareto Evolutionary Algorithm (SPEA) was introduced to find a set of approximate Pareto efficient solutions. SPEA uses a correction method for handling the constraints of combinatorial problems, resulting in an interruption of the evolutionary process. In this paper, we present an extension of the original algorithm, which keeps infeasible solutions in the population, avoiding a slowdown of the evolutionary process. The former correction method is thus no longer required. The resulting DPEA (distance-based Pareto evolutionary algorithm) uses a modified fitness function based on the Euclidean distance to the closest solutions in the Pareto front. The functionality of DPEA and its implementation are discussed. As a first result, DPEA performs better than SPEA on optimizing knapsacks with weaker constraints (knapsacks with a higher capacity).