Beam-GBAS for Multidimensional Knapsack Problems.

In order to improve the performance of Graph-Based Ant System (GBAS) to solve Multidimensional Knapsack Problem (MKP), a method combining GBAS with Beam Search (Beam-GBAS) is proposed. GBAS transforms the unordered information of the problem into ordered influence on ants. Beam search is used to increase the number of candidate paths obtained by GBAS and the optimal path will be obtained by comparing. The experimental results show that the proposed method not only improves the performance of searching the optimal solution, but also improves the stability compared with GBAS.