An object-oriented methodology for solving assignment-type problems with neighborhood search techniques

Because of its specificity, it is usually difficult to reuse computer code developed for a given combinatorial problem to deal with another one. We use the Object Oriented Programming methodology to derive general purpose software including four different neighborhood search techniques (descent method, tabu search, exchange procedure, simulated annealing) to deal with any assignment-type problem with a bare minimum of coding effort. The structure of the software even allows a more advanced user to play around with several parameters of these techniques and to modify the techniques to create specific variants.