Compression-coded genetic algorithms and its application in traffic control

The compression coding, generated by suitable compressing method, is applied in a genetic algorithm. Its length, together with its corresponding length of the non-compressed coding, can be altered. As one compression-code bit is ordinarily on behalf of several non-compressed bits, changes in any compression-code bit will create changes in several non-compressed ones. Therefore, the compression-coded genetic algorithm is helpful to improve the diversity of the model and the parallelism of the algorithms. In practice, according to the constrict condition and a self-adaptive decoding function, the compressed code genetic algorithm can avoid a premature convergence of the evaluation of multi-extremum in genetic algorithms in some degrees, therefore improve the search capacity of the global optimal solution. Through the multiobjective optimization by applying the compression coding genetic algorithms on the basis of the dynamic multi-driveway four-phase network model of urban traffic, the simulation result is confirmed to be feasible.