Solving the Multidimensional Multiple Knapsack Problem with Packing Constraints Using Tabu Search

This paper presents a methodology for solving military aircraft load- scheduling problems modeled as a multidimensional multiple knapsack problem, Because of the computational time associated with applying conventional algorithms to this type of problem, we employ tabu search to determine how much cargo a heterogeneous group of aircraft can carry. This study extends the previous work of Chocolaad in two areas. First, we modify Chocolaad's algorithms to solve the multiple (rather than the single) knapsack problem under the constraints he defmed for the Airlift Loading Problem. Second, we drop his assumption of a homogeneous group of aircraft. We validate our model by confirming its solutions with cargo loadmasters, and comparing the performance of our algorithm with the benchmark ALM.