Paper Conference

Proceedings of Building Simulation 2011: 12th Conference of IBPSA


Heuristic Vs. Meta-Heuristic Optimization For Energy Performance Of A Post Office Building

Won-Jun Suh, Cheol-Soo Park, Deuk-Woo Kim

Abstract: In this paper, application of heuristic and meta-heuristic to energy optimization of a post office building is presented. The target building was first optimized by a heuristic approach which was based on the expertise, experience and intuition of experts with the use of a whole building simulation tool, EnergyPlus. Then, the heuristic approach was compared to one of the metaheuristic approaches, Genetic Algorithm (GA), in terms of energy savings and the pros and cons of each approach. The meta-heuristic approach was completed in the MATLAB platform where EnergyPlus and GA are coupled. An M-script file was developed to automate execution of simulation runs (reading output files and writing input files) with GA. It should be noted that in this study, most design and simulation parameters were fixed and given by the client. Therefore, the simulationists were not allowed to make any significant change to the original design of the building. It is not surprising that GA is much better in finding a global optimum than the heuristic approach but it takes significant simulation run times and programming effort. The heuristic approach has advantages such as reflection of design context in decision-making and fast communication between stakeholders.
Pages: 704 - 711