Load balancing solution for heterogeneous wireless networks based on the knapsack problem

This paper proposes a practical approach to load balancing in heterogeneous wireless networks, which is based on the adaptation of the multiple knapsack problem to this specific scenario. Several algorithms capable of solving the knapsack problem are considered, namely a brute force, a dynamic programming and a greedy approach. The algorithms are evaluated by computer simulations in terms of the capability of finding the optimal solution and of computational complexity. The best algorithm identified, capable to generate near to optimal solutions while keeping low the computational complexity, is integrated into a load balancing mechanism intended to be used in real communication networks. The test scenario and the results obtained by practical experiments are presented and analyzed.