EFFICIENT DISTRIBUTION OF FLY ASH BASED ON ARTIFICIAL INTELLIGENCE MODEL: A CASE OF NTPC, FARAKKA (Paper 798)

Supply chain is a complex and dynamic system encompassing the behaviour of different participants. The paper attempts to develop a neuro-fuzzy model to capture the behaviour of supply chain, demand modelling and selection of best mode for disposal of fly ash from case study location, NTPC Farakka. Traditional models suffer from limitations which affect the accuracy of their forecasting. In this study, neuro-fuzzy model has been designed to predict the distribution pattern and mode selection. Primary objective of the study is to capture the firm’s behaviour, which influences mode choice and distribution. Four parameters namely, production capacity, travel cost, travel time and mode preference were considered to capture the behaviour of users. Three modes have been considered in this study namely road, rail and iwt. Output from the model predicts the flow of fly ash to different clusters through different modes. In case of cement industries, the base year demand was 68783MT/month while model prediction was 65568MT/month, having an error of only 4%, while in case of brick industries 10% variation was observed. Overall mode share obtained from the model based on individual cluster behaviour: 36% by road, 45% by rail and remaining 19% by IWT. Thus, an attempt was further made to explore the potential opportunities of transport system that best suits each cluster. Best mode selection was carried out by incorporating both, the obtained mode choice behaviour of firms as well the economical and sustainable aspects.