A novel artificial bee colony algorithm for HVAC optimization problems

Abstract

The strategy of simulation-optimization is popular in handling the design and energy management of centralized HVAC systems. An effective optimization algorithm is paramount to obtain the optimal design and operation of HVAC systems. Due to the multi-dimensional and highly constrained nature commonly found in HVAC problems, traditional optimization methods are generally not applicable. Instead, a metaheuristc algorithm, e.g., the genetic algorithm, is a good choice to achieve these purposes. In recent years, the paradigm of the artificial bee colony is emerging in the field of evolutionary computation. The artificial bee colony optimizes the problem through the foraging behavior of honey bees. This article presents a novel approach developed from this paradigm, called the one-position inheritance artificial bee colony algorithm, for handling HVAC system optimization. Through two HVAC example problems, it is found that the one-position inheritance artificial bee colony can more effectively and efficiently search the optimal solutions against the genetic algorithm and original artificial bee colony. This suggests that the one-position inheritance artificial bee colony is a good optimization algorithm to solve other HVAC engineering problems with a multi-dimensional, mixed discrete-continuous, and highly constrained search landscape.