Dynamic ant colony algorithm based on knowledge base

The ant colony algorithm (ACA) has the limitation of stagnation and is easy to fall into local optimums. Therefore, the characteristics of the algorithm are researched, and a dynamic ant colony algorithm (DACA) based on knowledge base is proposed. The knowledge base consists of algorithm, rule, and case knowledge. The qualitative or quantitative algorithm parameter, parameter choosing method, and history data are saved to the knowledge base. DACA generates an initial state, dynamically adjusts the parameter based on the knowledge base and model state, and chooses the parameter through roulette wheel selection. DACA can quickly converge to the global optimization solution without the influence of random search process. Eil51 and CHN144 are solved by DACA and other algorithms. Result shows that DACA is the best in optimization performance, time performance, and robustness.