Green Roof Optimization Using Multi Objective Optimization with Genetic Programming Based Artificial Neural Network

In recent years, the greenhouse gases are rising increasingly and the extreme usage of vestige energy resources helped in the green spaces development including green roof systems, which is an important focus in urban areas. In existing, green roof designs are optimized with different methods such as Fuzzy framework-based optimization. However, the framework is a single object based optimization so, the framework struggle to handle multi conflict objects. To overcome this issue, Multi Objective Optimization with Genetic Programming based Artificial Neural Network (MOO-GPANN) is proposed to optimize the green roof design. Initially, the input data is taken from NYC Green Roof Footprints (NYC-GRF) dataset and further processed into feature extraction with fuzzy if-then rule. Then, the optimal features are selected by using Multi Objective Optimization (MOO) based on ranking diversity. Finally, the prediction done by using Genetic Programming based Artificial Neural Network (GPANN). From the results, the proposed MOO-GPANN model gave better results than existing Adaptive Neuro-Fuzzy Inference System (ANFIS) by providing results for the objective functions including energy consumption (PMV), Heating Load (HL), Cooling Load (CL) and comfort levels (CFL) respectively.