Abstractachine learning methods can be used to help design energy-efficient buildings reducing energy loads while maintaining the desired internal temperature. They work by estimating a response from a set of inputs such as building geometry, material properties, project costs, local weather conditions, as well as environmental impacts. These methods require a training phase which considers a dataset drawn from selected variables in the problem domain. This paper evaluates the performance of four machine learning methods to predict cooling and heating loads of residential buildings. The dataset consists of 768 samples with eight input variables and two output variables derived from building designs. The methods were selected based on exhaustive research with cross validation. Four statistical measures and one synthesis index were used for the performance assessment and comparison. The proposed framework resulted in accurate prediction models with optimized parameters that can potentially avoid modeling and testing various designs, helping to economize in the initial phase of the project. Keywords: Energy efficiency. Heating and cooling loads. Machine learning. Resumo

The constrained optimization problems are very common in the engineering field. For instance, in civil, aeronautical, mechanical engineering and so on, this type of problem is largely used to find the best designs of structures leading to a structural optimization problem to be solved. Commonly, these problems consist in to find structures with the minimum weight, subject to a set of constraints such as allowable stress, displacements, natural frequencies of vibration and stability criteria. Besides the traditional optimization methods, consolidated through the decades, the evolutionary algorithms, in general inspired by natural phenomenona, have been playing an important role showing robustness to solve this kind of problem. In 2005, the artificial bee colony algorithm (ABC), inspired by the foraging of bee colonies, was proposed to solve multimodal and multidimensional optimization problems. This paper proposes, analyzes and discusses the coupling of ABC to variants of an adaptive penalty method, handling the constraints, to solve traditional problems of truss structural optimization. The results obtained are compared with the literature showing that the proposed strategy can be efficient and competitive.

Purpose The purpose of this paper is to evaluate the performance of social spider algorithm (SSA) to solve constrained structural optimisation problems and to compare its results with others algorithms such as genetic algorithm, particle swarm optimisation, differential evolution and artificial bee colony. Design/methodology/approach To handle the constraints of the problems, this paper couples to the SSA an efficient selection criteria proposed in the literature that promotes a tournament between two solutions in which the feasible or less infeasible solution wins. The discussion is conducted on the competitiveness of the SSA with other algorithms as well as its performance in constrained problems. Findings SSA is a population algorithm proposed for global optimisation inspired by the foraging of social spiders. A spider moves on the web towards the position of the prey, guided by vibrations that occur around it in different frequencies. The SSA was proposed to solve problems without constraints, but these are present in most of practical problems. This paper evaluates the performance of SSA to solve constrained structural optimisation problems and compares its results with other algorithms such as genetic algorithm, particle swarm optimisation, differential evolution and artificial bee colony. Research limitations/implications The proposed algorithm has no limitations, and it can be applied in other classes of constrained optimisation problems. Practical implications This paper evaluated the proposed algorithm with a benchmark of constrained structural optimisation problems intensely used in the literature, but it can be applied to solve real constrained optimisation problems in engineering and others areas. Originality/value This is the first paper to evaluate the performance of SSA in constrained problems and to compare its results with other algorithms traditional in the literature.

No abstract

No abstract