A Comprehensive Review on Evolutionary Algorithm Solving Multi-Objective Problems

Multi-Objective Optimization of Ultra-Low Energy Housing in Hot Summer Cold Winter Climate Zone of China Based on a Probabilistic Behavioral Model

“…In the field of building performance simulation, various optimization methods are used to solve multi-objective problems. There are two common approaches to the problem of optimizing buildings with the goals of reducing energy consumption or improving the indoor thermal environment, (1) the classical weighted sum approach [34] and (2) the Pareto-dominance approach [35]. The classical weighted sum approach converts MOO into a single scalar objective problem by attaching a corresponding weighting (e.g., weighting and algorithms) to individual objectives based on mathematical principles [36]; this method does not provide information about the mutual interference between different sub-targets.…”

“…The classical weighted sum approach converts MOO into a single scalar objective problem by attaching a corresponding weighting (e.g., weighting and algorithms) to individual objectives based on mathematical principles [36]; this method does not provide information about the mutual interference between different sub-targets. The Pareto-dominance approach employs stochastic rules to find the set of non-dominated solutions in the entire space of feasible decision variables, i.e., Pareto solutions [35], optimizing all objectives simultaneously and providing a set of non-obvious optimal solutions, facilitating decision makers in choosing the optimal solution according to different preferences [36]. Therefore, the Pareto dominance method is increasingly used in the field of multi-objective optimization.…”

Optimization Based on Computational Fluid Dynamics and Machine Learning for the Performance of Diffuser-Augmented Wind Turbines with Inlet Shrouds

“…Therefore, the most common solution concept is to compute a set of Pareto optima, solutions that cannot be improved in one objective without accepting a worsening in others, and then let a decision-maker select the final solution based on their preference [63]. As a mainstream method for solving MOPs, the development and application of evolutionary algorithms (EAs) has attracted thousands of researchers since the 1950s [64]. EAs profit from their general ability to work with sets of solutions, are the standard approach to MOPs, and have many successful applications [63].…”

“…EAs profit from their general ability to work with sets of solutions, are the standard approach to MOPs, and have many successful applications [63]. The NSGA-II algorithm, a model initially proposed by Deb et al [65] in 2002, is considered the most prominent multi-objective EA [63,66] with the most popular GA framework [64], and has served as a powerful decision-space exploration engine, based on GA, used to solve MOPs [67]. So far, it has been cited more than 50,000 times on Google Scholar [68] and is becoming one of the most widely used algorithms for solving MOPs in various applications in different fields [12,[68][69][70].…”