MinOrder optimization in FAP with nogood-tabu search

The problem considered in this paper consists in defining an assignment of frequencies to radio links between transmitters/receivers which minimize the global interference inside the network. This problem can be modeled as a Constraint Satisfaction Problem CSP. It is NP-hard and several results have been reported on techniques for solving it optimally. We applied to this version of the frequency assignment problem an original hybrid method, which combines constraint propagation techniques and Tabu search. The method is based on finding and storing the variable/value associations that lead to infeasible solutions. It allows the algorithm to dynamically manage the cuts of branches and the access to the decision variables. It is a learning system which progressively extends the algorithm knowledge on the problem and increases its effectiveness. Computational results obtained on a number of standard problem instances prove the effectiveness of the proposed approach.