In this paper, a new meta-heuristic algorithm called modified social optimisation algorithm (MSOA) is proposed to solve reactive power control problem and similar minimisation problems. MSOA which is a double-phase strategy and has been inspired by the behaviour of individuals in an ideal society is the enhanced version of social optimisation algorithm (SOA). The contribution of this paper lies in the enhancements made to SOA. MSOA utilises changeable personal choices and updating coefficients. Moreover, the initial individuals in the proposed algorithm are evenly distributed. The proposed algorithm is capable of searching a larger space to find the optimal solutions. The convergence of the algorithm to local minima is significantly avoided through designating larger ranges for the control parameters of the algorithm and also replacing a large percentage of the worst individuals by randomly-distributed individuals. The method has been tested on 23 well-known functions and the results have been compared with the others algorithms in the literature. In 87% of the benchmark functions, MSOA gives better or similar values in comparison with well-known algorithms. By employing our proposed approach in reactive power control problem, the transmission losses reduce by 16% for IEEE 57-bus system and 8.4% for IEEE 14-bus system.
