Ant colony optimization inspired resource allocation for multiuser multicarrier systems

In this paper, we present an ant colony optimization (ACO) based Bayesian learning which optimizes the transmission energy efficiency for a multi-user multi-carrier communication system. The ACO algorithm is inspired by the swarm behavior of ants — sharing out the work and helping one another, which is a kind of Bayesian learning. The resource allocation problem in a multi-user multi-carrier system is reformulated by utilizing pheromone trail and heuristic information in ACO. As the multicarrier multiple access resource allocation problem is NP-hard, many greedy, yet suboptimal, algorithms have been studied. The swarm-based ACO algorithm makes decision stochastically to reduce chance of being trapped in a local optimality and to draw near the optimal solution. In this work, the proposed ACO method uses the strength of masses to find candidate routes and helps to approach optimal solution while satisfying all the system constraints. The classical meta-heuristic ACO algorithm is reformulated and is combined with rate relaxation convex optimization to achieve the optimization goal. The proposed method is numerically verified in simulations that achieves near optimal energy efficiency as compared with the exhaustive search method.