This paper presents a heuristic approach to generate initial feasible solutions for the Unit Commitment (UC) problem in electric power generation. The Chemical Reaction Optimization (CRO) algorithm is implemented to solve this problem. Multiple generator constraints and system constraints are considered. We also program the binary PSO and the Elite PSO (EPSO) for comparison. The proposed heuristic approach is combined with the three optimization algorithms to form H-CRO, H-PSO and H-EPSO. We test the performance of all algorithms on the standard 10-unit system. Simulation results show that the heuristic can improve the performance and CRO provides better convergence than the two PSO algorithms. H-CRO is also tested on a 20-unit and 100-unit system to show its capability. The results provided in this paper suggest that the proposed heuristic approach is a better alternative for solving the UC problem. CRO also has its advantage in optimizing UC problems.