The latest efficient MILP model is adapted to general cases. Improvements on an existing heuristic method are proposed. Global way to update a new generation. Consideration of the asymmetry of the problem. Effectiveness of the improved methods is verified by case study.