Abstract
SummaryDynamic multi-objective optimization (DMO) is one of the most challenging class of multi-objective optimization problems where the objective function(s) change over time and the optimization algorithm is required to identify the corresponding Pareto optimal solutions with minimal time lag. DMO has received very little attention in the past and none of the existing multi-objective algorithms perform too well on the set of newly proposed DMO test problems. In this chapter, we introduce a memetic algorithm (MA) to deal with such class of problems. The memetic algorithm employs an orthogonal epsilon-constrained formulation to deal with multiple objectives and a sequential quadratic programming (SQP) solver is embedded as its local search mechanism to improve the rate of convergence. The performance of the memetic algorithm is compared with an evolutionary algorithm (EA) embedded with a Sub-EA on two benchmarks FDA1 and modified FDA2. The memetic algorithm consistently outperforms the evolutionary algorithm with Sub-EA for both FDA1 and modified FDA2 for all problem sizes. A variational study on the effect of the number of SQP iterations on the behavior of MA is included to highlight the benefits offered by MA for such class of problems.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have