Multi-objective life cycle utility optimization

The optimal design of a structure is usually achieved through optimizing life cycle utility using a single objective optimization approach. The Life cycle utility function aggregates all benefit, capital cost and life time costs. Another approach is multi-objective optimization (MOOP) which allows us to explore trade-offs between conflicting cost functions. The MOOP concept allows the decision maker to select the most appropriate design alternative depending on his/her preference. The optimal pool of solutions is represented by the Pareto-optimal front. One of these solutions matches the optimal design obtained using single objective optimization (SOOP). Three classical methods, namely, weighted sum, weighted-Tchebycheff and epsilon-Constraint methods together with an evolutionary algorithm (genetic algorithm GA) are investigated.. It is found that multi-objective life cycle utility optimization can be achieved effectively using either the classical methods or GA. However, GA tends to involve less computational effort. An application and comparison between methods are presented.