Differential evolution (DE) algorithm is widely used optimization method. A population of popular adaptive DE variant (jSO) is distributed to several sub-populations to increase its efficiency. Three crucial control parameters of proposed migration model (mSO) are experimentally studied. Real-world problems of CEC 2011 are used as a benchmark to show the efficiency of mSO versions in solving the real problems compared to original jSO. Presented results show good performance of mSO especially in case of limited time. Higher performance of mSO is achieved rather for lower or middle values of the migration control parameters.