A New Ant Colony Optimization Applied for the Multidimensional Knapsack Problem

This paper proposes a Binary Ant System (BAS), a new Ant Colony Optimization applied to multidimensional knapsack problem (MKP). In BAS, artificial ants construct the solutions by selecting either 0 or 1 at every bit stochastically biased by the pheromone level. For ease of implementation, the pheromone is designed specially to directly represent the probability of selection. Experimental results show the advantage of BAS over other ACO based algorithms. The ability of BAS in finding the optimal solutions of various benchmarks indicates its potential in dealing with large size MKP instances.