keywords: Weather, forecasting, fuzzy logic, optimization, rules-lists, accuracy, redundancy, genetic algorithms
The numerical weather prediction relies strongly on the precipitation forecast because of the use to civil protection agencies, enterprises, daily activities of people around the world, and the reduction of economic and social damages. However, there is the need to evolve better and more accurate parameterization of physical processes in order to improve on the outcomes of forecasts generated. This seminar paper discusses the weather forecasting issues, methods and gaps of existing solutions. In particular, evolutionary computing and fuzzy logic techniques are being investigated for the purpose of developing an effective model that could guarantee better accuracy and reliability of outcomes when applied for weather uncertainty problems in real-world situation. It was found that, the fuzzy logic approach has low accuracy, which needs to be improved with rule-lists optimization. The genetic algorithms hold promise in overcoming these tasks, by providing important information on weather and state of the atmosphere in certain places and periods through the application of fuzzy logic and genetic algorithms for improved outcomes of weather forecasts.