Stochastic Tabu Search strategy and its global convergence

Tabu Search (TS) is a metaheuristic that guides a local heuristic search procedure to explore the solution space beyond local optimality. It has achieved widespread successes in solving practical optimization problems. This paper proposes a stochastic TS strategy for discrete optimization and makes an investigation of its global convergence. The strategy introduces the Metropolis criterion and annealing process of simulated annealing technology into a general framework of TS. It has been proved that the strategy converges asymptotically to global optimal solutions, and satisfies the necessary and sufficient conditions for global asymptotical convergence. Furthermore, it produces a higher convergent rate than simulated annealing algorithm.