Using software agents for weather forecasting from multiple perspectives

Abstract

Weather forecasting has become important for many sectors such as agriculture, fisheries and transportation. A reliable weather forecast report is of great importance for those sectors to plan their activities by mitigating difficulties they face from bad weather conditions. Although there are a number of weather forecasting methods available, their false positive rates are considerably high. This study presents the use of Multi Agent technology to ensemble weather forecasting results independently by artificial neural network and fuzzy logics. We define two agents to operate on artificial neural network and fuzzy logic solutions that start negotiation & deliberation to produce a more rational forecasting. This process enables sharing findings of the two agents and improves the overall quality of the forecasting.