Performance Assessment of Algorithms for Building Energy Optimization Problems with Different Properties

Abstract

Assessing the performance of algorithms in solving building energy optimization (BEO) problems with different properties is essential for selecting appropriate algorithms to achieve the best design solution. This study begins with a classification of the properties of BEO problems from three perspectives, namely, design variables, objective functions, and constraints. An analytical approach and a numerical approach are proposed to determine the properties of BEO problems. Six BEO test problems with different properties, namely, continuous vs. discrete, convex vs. non-convex, linear vs. non-linear, uni-modal vs. multimodal, and single-dimensional vs. multi-dimensional, are composed to evaluate the performance of algorithms. The selected optimization algorithms for performance assessment include the discrete Armijo gradient, Particle Swarm Optimization (PSO), Hooke-Jeeves, and hybrid PSO and Hooke-Jeeves. The assessment results indicate that multimodality can cause Hooke-Jeeves and discrete Armijo gradient algorithms to fall into local optima traps. The convex, non-convex, linear and non-linear properties of uni-modal BEO problems have little impact on the performance behavior of the algorithms. The discrete Armijo gradient and Hooke-Jeeves are not recommended for solving discrete and multi-dimensional BEO problems. automatically adjust designs, has emerged [3]. It combines optimization techniques with building energy simulation tools and relies on optimization algorithms to create new designs according to pre-defined optimization objectives and simulation results [4]. Thus, optimization algorithms are the key to the BEO workflow, and the effectiveness and efficiency of this technique significantly depend on the performance of the algorithms.There are various algorithms for solving optimization problems in many science and engineering fields. The optimization algorithms commonly used in BEO can be generally divided into three groups, namely, hybrid algorithms, direct search algorithms and heuristic algorithms [5]. As shown in some reviews [6][7][8], evolutionary algorithms are the most popular algorithms, accounting for about 60%. The genetic algorithm (GA) [9] and its variations such as the non-dominated sorting genetic algorithm II (NSGA-II) [10] are typical examples of evolutionary algorithms. The Hooke-Jeeves algorithm [11] is a representative of direct search algorithms. A hybrid algorithm [12] is an algorithm that combines two or more other algorithms into a hybrid operation so that the overall algorithm performs better than the individual ones.Although a number of algorithms are currently used in BEO, their performance, i.e., in terms of effectiveness and efficiency, can vary and at times significantly. The reasons for the difference in performance of algorithms are two-fold. First, BEO problems have different properties, e.g., linear or non-linear, single-dimensional or multi-dimensional, uni-modal or multimodal, etc. Second, the performance of a particular algorithm is closely linked with the properties of...

A Review of Performance-Oriented Architectural Design and Optimization in the Context of Sustainability: Dividends and Challenges

“…The effectiveness and efficiency of conventional scenario-by-scenario calculations and manual control of multiobjective optimization methods are limited by the breadth of design and the professional skills of the participants [51]. Advances in artificial intelligence technology have opened up new possibilities in the design process and the way of interaction.…”

“…On the contrary, it causes unnecessary energy consumption. A lot of correlation knowledge emerged in this survey to make up for the data gap at this stage [40,42,48,51,55,62,65]. Changes in the data of reference models for different climate information are usually not recorded or tracked.…”

Study on the Optimum Design of a Ground Heat Pump System Using Optimization Algorithms

“…There are various optimization algorithms, including gradient algorithms, pattern search algorithms, and genetic algorithms. Si et al [14] applied various optimization algorithms to the design of buildings. They analyzed the characteristics of design elements, such as the direction, area ratio, and wall conductivity, which are passive elements, and then performed optimization to obtain values that minimized energy consumption.…”

“…The parameters of the discrete Armijo gradient algorithm were set by referring to the paper that used this optimization algorithm [14].…”

“…The PSO algorithm did not need a large capacity because it was simple and only required a short calculation time. The possibility of the occurrence of a local optimization was also slim, and could be used in various areas because it could be applied to both continuous variable and discrete variable problems [14]. In this study, the number of particles and the number of generations were set as defaults, the number of particles was 8, the number of generations was 16, and the PSO algorithm was optimized.…”