Energy management and design of centralized air-conditioning systems through the non-revisiting strategy for heuristic optimization methods

A directional mutation operator for differential evolution algorithms

“…Initially five solutions are randomly created. Suppose they are G 1 = {x 1,g , · · · , x i,g , · · · , x 5,g }: [4,4] T . Accordingly, their function values are…”

“…In EAs, the function to be optimized is often called fitness functions; the domain of variables called search space; a feasible solution called individual and the function value of a feasible solution called fitness or fitness value. In practice, EAs have been applied to many fields such as engineering design [3], energy management [4], financial strategies [5] and computer vision [6] etc. These applications justify the usefulness of EAs.…”

Analysis of (1+1) Evolutionary Algorithm and Randomized Local Search with Memory

“…This gain can be enormous for applications that involve expensive or time-consuming simulations to evaluate the fitness of a point. Examples can be found in works by Fong et al (2008Fong et al ( , 2010, which use simulation programs to evaluate the design of a heating, ventilating, and air-conditioning (HVAC) system. Clearly, using simulation to evaluate the fitness of a particular parameter setting is time-consuming, and it is commented that such a simulation-optimization approach has increasingly been applied to the study of HVAC systems (see Fong et al, 2008 and references therein).…”

A thermal resistance-based method for the optimal design of central variable water/air volume chiller systems

“…These methods help to design the structural parameters for specific systems, whereas they do not perform well in other systems different from the one with the empirical or experimental data. Meanwhile, many algorithm-based methods have also been raised, such as the particle swarm optimization method [18] and the genetic algorithm (GA) method [19]. These algorithms could provide the results close to the optimal ones in the situation of lacking some information of a system [20], whereas they require plenty of iterations.…”

“…One focuses on the system models, including the bi-objective optimization model [2], the empirical/ experimental-data-based model [7,10,11,21], the detailed physical dynamic model [22], the transient thermal model [23], and the simulated model [24,25]. Two mainly aims at the mathematical algorithms, such as the neural network (NN) algorithm [2,26,27], the sequential quadratic programming algorithm [28], the genetic algorithm (GA) [4,19,24], the adaptive neuro-fuzzy inference algorithm [29], the support vector data description (SVDD) [30] and the enumeration algorithm [31]. The last pays attention to the control technologies combined with the system models or the mathematical algorithms, involving the temperature/flow-rate/pressure control strategy together with the wavelet NN algorithm [32], the feedback control strategy with the NN and the GA [26], the on-line control strategy along with the dynamic models and the GA [20] and the simulated model with the GA [24].…”