Applying ant colony optimization to dynamic binary-encoded problems

Ant colony optimization (ACO) algorithms have proved to be able to adapt to dynamic optimization problems (DOPs) when stagnation behaviour is addressed. Usually, permutation-encoded DOPs, e.g., dynamic travelling salesman problems, are addressed using ACO algorithms whereas binary-encoded DOPs, e.g., dynamic knapsack problems, are tackled by evolutionary algorithms (EAs). This is because of the initial developments of the algorithms. In this paper, a binary version of ACO is introduced to address binary-encoded DOPs and compared with existing EAs. The experimental results show that ACO with an appropriate pheromone evaporation rate outperforms EAs in most dynamic test cases.