Use of a Genetic Heritage for Solving the Assignment Problem with Two Objectives

The paper concerns a multiobjective heuristic to compute approximate efficient solutions for the assignment problem with two objectives. The aim here is to show that the genetic information extracted from supported solutions constitutes a useful genetic heritage to be used by crossover operators to approximate non-supported solutions. Bound sets describe one acceptable limit for applying a local search over an offspring. Results of extensive numerical experiments are reported. All exact efficient solutions are obtained using Cplex in a basic enumerative procedure. A comparison with published results shows the efficiency of this approach.