MGA-ACO Based control system for the large-scale solid garage

A new encoding method and a novel hybrid MGA-ACO (modified genetic algorithm and ant colony optimization) were proposed for the parking scheduling problem in large solid garages. Combining with the 3-dimensional modal of the garages, the new form of encoding can easily reflect the real state of the garages, which makes the mathematical tools be more applicable to the problem. In the MGA-ACO algorithm, ACO offers a critical advantage of local searching, while GA considers a global perspective operating on the complete population from the very beginning. Therefore, ACO and GA can nullify each others´ drawbacks when hybridized. The best solution may generated by either ACO or GA, that is ACO can utilize the GA´s cross-over and mutation operations. The experimental results showed the superiority of MGA-ACO to MGA and ACO, both in stability and convergence.