This paper examines recent developments of service restoration in distribution systems (DS). Recent approaches and challenges for service restoration are compared and analyzed. Service restoration is a multi-constraint, multi-objective, combinatorial, non-linear optimization problem. Its main aim is to maximize the number of priority loads restored, while minimizing the number of switching operations, within the shortest time interval. This paper introduces Hybrid Multi-Agent System Approach for service restoration, which uses Distributed Generators (DG) and Electric Vehicles (EVs). EVs are basically batteries, which are charged at low-load conditions and transmit required amount of energy at peak-load conditions back to the grid. Before applying this system in real cases, optimal positions of DGs must be found. In order to do so OpenDSS Distribution network simulator was used for IEEE 123 and 30 Bus System. After that, RM algorithm was developed and implemented under IEEE 30 Bus System. The main goal of this algorithm is to determine the most optimal island ranges according to the given objective functions and constraints. Proposed RM algorithm demonstrated quite promising results as it fulfill all preset requirements.