Particle Swarm Optimization (PSO), which has attracted a great deal of attention as a global optimization method in recent years, has the drawback that continuous search based on its excellent dynamic characteristics cannot be performed stably until the end of computation due to its very strong tendency to convergence. In this paper, we propose a "Repetitive Search Guideline" which differs from the common guidelines in the improved methods which have since been proposed and by which the continuous search in PSO is achieved without losing PSO's excellent dynamic characteristics due to repetitive search in a promising area where the objective function values are expected to be small. We consider four improved methods based on the proposed guidelines, then confirm their effectiveness by application to 100-variable multipeaked benchmark problems.