This research aims to find an optimization method that is applicable to the real-time control of offsets and cycle lengths and that is better than Genetic Algorithm (GA) from a viewpoint of calculation efficiency. For this purpose, some methods are selected from metaheuristics that is a generic name of stochastic search methods like GA and that intends to find a global optimal solution. Those selected methods are compared through application to offset optimization. As a result, ILS is adopted as a best search method. After that, ILS is applied to sequential adjustment of offset and cycle length. It was found that ILS was applicable to real-time control strategy of signal coordination.