Modern engineering problems, such as aircraft or automobile design, are often composed by a large number of variables that must be chosen simultaneously for better design performance. Normally, most of these parameters are conflicting, i.e., an improvement in one of them does not lead, necessarily, to better results for the other ones. Thus, many methods to solve multi-objective optimization problems (MOP) have been proposed. The MOP solution, unlike the single objective problems, is a set of non-dominated solutions that form the Pareto Curve, also known as Pareto Optimal. Among the MOP algorithms, we can cite the Firefly Algorithm (FA). FA is a bio-inspired method that mimics the patterns of short and rhythmic flashes emitted by fireflies in order to attract other individuals to their vicinities. For illustration purposes, in the present contribution the FA, associated with the Pareto dominance criterion, is applied to three different design cases. The first one is related to the geometric design of a clamped-free beam. The second one deals with the project of a welded beam and the last one focuses on estimating the characteristic parameters of a rotary dryer pilot plant. The proposed methodology is compared with other evolutionary strategies. The results indicate that the proposed approach characterizes an interesting alternative for multi-objective optimization problems.
Multi-Objective Optimization Firefly Algorithm Applied to (Bio)Chemical Engineering System Design
