Finding Tradeoffs by Using Multiobjective Optimization Algorithms

Abstract

The objective of the present study is to demonstrate performances of Evolutionary Algorithms (EAs) and conventional gradient-based methods for finding Pareto fronts. The multiobjective optimization algorithms are applied to analytical test problems as well as to the real-world problems of a compressor design. The comparison results clearly indicate the superiority of EAs in finding tradeoffs.

Aerodynamic optimization via multi-objective micro-genetic algorithm with range adaptation, knowledge-based reinitialization, crowding and ε-dominance

“…After letting these to act on the population for sufficiently many generations, the evolution eventually gets to the optimum. In the case of a single objective (SOGAs), this would be a single member of the population, whereas in the multi-objective case (MOGAs), a set of nondominanted or Pareto-optimal tradeoff solutions appears, lying on the global Pareto front of the design space [6]. To be a bit more precise: a decision vector x 1 dominates a decision vectorx 2 if: f i ðx 1 Þ 6 f i ðx 2 Þ; 8i ¼ 1; .…”

Multi-objective optimization of an axial compressor blade

“…In multi-objective optimizations of turbomachinery blades, efficiency, total pressure, static pressure, pressure loss, weight, stress, etc. are used as objectives, and variables related to camber profile and/or stacking line of blade are employed as design variables [15][16][17][18]. A multi-objective optimization problem consists of many optimal solutions called Pareto-optimal solutions; therefore, a designer's aim is to find as many optimal solutions as possible within the design range.…”

Diversity Maintenance Mechanism for Multi-Objective Genetic Algorithms Using Clustering and Network Inversion

“…However, in real-world problems, such as large-scale design problems [3,4], the reduction of evaluation calls becomes an essential issue due to their high computational cost. Two major approaches have been proposed for this issue: the response surface method [5][6][7] and search using small population size [8] (SSP strategy). In this paper, an effective SSP mechanism is discussed.…”