Table 1: Multiple objective optimization algorithms comparison.
Optimization Method |
Advantages |
Disadvantages |
Weighted Sum Method |
· Simple |
· Difficult to set the weight factors · Not applicable to non-convex problems |
ε-Constraint Method |
· Applicable to both convex and non-convex problems |
· The ε vector has to be chosen carefully |
Weighted Metric Method |
· Guarantees finding all Pareto-optimal solutions with ideal solution |
· Requires min and max objective values · Requires that ideal solution can be found by independently optimizing each objective function |
NSGA-II & NCGA |
· Non-penalty constraint handling · Fast and efficient convergence · Search in a wide range and handle problems that start with non-feasible solutions |
· Complexity |
SPEA2 |
· Once a solution on Pareto-optimal front is found, it is stored in the external population · Advantages from NSGA-II |
· Requires a balance between the regular population size and the external population size |
MOPSO |
· Simple |
· Worse than GAs in finding pareto solutions |
MOSA |
· Fast |
· Worse than GAs in finding pareto solutions |