Building Information Modeling using Hardware Genetic Algorithms with Field-Programmable Gate Arrays

Genetic algorithms (GAs) have found many applications in various fields such as physics, signal processing, artificial intelligence and recently construction engineering management. For a long time, GAs are usually criticized to be time-consuming, making it unpractical for real-time applications. This paper presents a new technique which can be used: (1) to automate construction activities, and (2) to improve building information modeling which has become an attractive research topic around the world. Different from the generic GA techniques employed in the literature, this paper proposes a new GA using hardware with field-programmable gate arrays. The proposed technique is shown to improve speed and lessen computational power. Hardware implementation of GA using static random access memory-based field-programmable gate arrays with synthesizable very hardware description language coding is introduced. Detailed analyses on the field-programmable gate arrays are given which show that it is suitable for real-time applications. As a result, GA is modified so that it can be implemented in series and parallel which can greatly improve computational hardware performance. Configuration of parallelization is available with a peripheral component interconnect interface, which further helps to form a fast optimization tool for real-time applications. The ultimate goal of this paper is thus to design an effective GA technique which can be employed to support building information modeling and to effectively automate critical processes in construction projects.