MICROCANONICAL OPTIMIZATION APPLIED TO THE TRAVELING SALESMAN PROBLEM

Optimization strategies based on simulated annealing and its variants have been extensively applied to the traveling salesman problem (TSP). Recently, there has appeared a new physics-based metaheuristic, calle d the microcanonical optimization algorithm (mu O), which does not res ort to annealing, and which has proven a superior alternative to the a nnealing procedures in various applications. Here we present the first performance evaluation of mu O as applied to the TSP. When compared t o three annealing strategies (simulated annealing, microcanonical anne aling and Tsallis annealing), and to a tabu search algorithm, the micr ocanonical optimization has yielded the best overall results for sever al instances of the euclidean TSP. This confirms mu O as a,competitive approach for the solution of general combinatorial optimization probl ems.