In this study, multi-verse-meshfree (MV-MF) simulation-optimization model and its multi-objective version (MOMV-MF) were developed to find the appropriate groundwater remediation strategy by PAT method. These models compare with particle swarm meshfree (PS-MF) simulation-optimization model and its multi-objective version (MOPS-MF). In the one-objective optimization problem, the objective function was defined as minimizing the mean of contaminant concentration by pumping at a constant rate, while in the two-objectives problem, the pumping rate was variable and the mean of pumping rate from all the wells is minimized. Finally, minimizing the remediation period added to set of objective functions as third objective function. The results indicated that locating the pumping wells in the path of the contaminant flow and close to the source improves the efficiency of the PAT system. In one-objective optimization problems, the results indicated that the MV-MF and PS-MF models have same performance with 2.68033 and 2.68035, respectively. But PSO algorithm is trapped in local optimum. Besides, the MVO algorithm, with its special theory of traveling objects between two universes, can easily move from one part of the search space to another. In the two-objectives problem, the results indicated that the number of solutions in the pareto-front in MOMVO with 8 solution was higher than MOPSO with 4 solution. In three-objectives problem, the results of MOMVO and MOPSO algorithm indicated that by changing the PAT policy, the 3 years (1095 days) remediation period can be reduced to 525 and 535 days, respectively. Also, one of the main findings of this research was good performance of MVO algorithm, because of the special way the algorithm works in transport the objects between different universes. The objects are nodes number and they can easily move between different universes but nodes number change in PSO algorithm by velocity vector.