Cost and Risk Prediction in Road Transportation of Hazmat by ANFIS Trained with PSO, FA, HBBO and ICA

Abstract

This paper proposes adaptive neuro-fuzzy inference system (ANFIS) to predict the risk with its aggregated cost (CR) of an accident in road transportation of hazardous material, the aim is to provide a more accurate and reliable data for the safety of transportation. The determination risk index by the conventional methods such as Risk graphs and deterministic approaches may lead to imprecise values due to the uncertainties, in both parameters and models. The proposed technique is a hybrid schema, which combines the main advantageous of fuzzy logic (address uncertainties) and neural network (learn from a given data). In other hand our study seeks to tune the parameters of the proposed model by particle swarm optimization (PSO), firefly algorithm (FA), imperialist competitive algorithm (ICA) and human based-behavior optimization (HBBO) and hence optimize the performance of ANFIS. The simulation result of this work and the comparative analysis shows that ANFIS yield height performance and the ANFIS-PSO was the outstanding one in the training phase, while ANFIS-FA gives better results in the testing process.