AES algorithm for dynamic knapsack problems in capital budgeting

Dynamic knapsack problems can model many economic phenomenon, such as capital budgeting. An agent based evolutionary search algorithm (AES) is proposed to solve a set of DKP problems which are generated according to dual mapping mechanism. Existing in a grid-like environment, all agents shall compete with their neighborhood to enhance their energy and acquire knowledge through Population-Based Incremental Learning. At the same time, triggered random immigration scheme is introduced into the algorithm to maintain the diversity of the population. Simulation results on a set of dynamic knapsack problems and t-test showed that AES with triggered random immigration scheme can obtain a better performance than several other similar genetic algorithms.