Modeling simulation and Optimization of cash management Using soft computing techniques for Banking applications

Abstract

Cash forecasting in banking operation is essential for budgeting, newlinemanaging the cash flow and maintaining the optimal cash balance. The newlinefinance officer must have the knowledge about the cash requirement without newlineaffecting the routine transaction. The bank plays a major role in the global newlinemarket to provide the effective customer service. Banks provide services for newlineboth the investors as well as the depositors to improve the economy of our newlinecountry. newlineCash forecasting is necessary to hold the sufficient cash for the newlinecustomers to make use of it during the ordinary days, festive days, salary day newlineand holidays. Banks have challenges to forecast the cash requirement with newlinebetter accuracy for maintaining the right amount of cash as well as to avoid newlinethe excess cash. Hence, there is a need to develop an efficient cash newlinemanagement model for banking operation. The present study is aimed at newlinedeveloping an optimized cash management model using computational newlineintelligence to reduce the liquidity risk. The conventional methods used by newlinebanks and other micro financial organization were statistical models such as newlineHoltz Model, Winter s Model and Moving average model. The statistical newlinemethods used by banks are time series methods or seasonal cash forecasting newlinemethods. To improve the accuracy of the cash forecasting process and the newlinefuture cash requirement can be determined using soft computing techniques newline.The soft computing based cash forecasting system is designed to reduce the newlineerror between the forecast value and the actual value. Hence, there is a need newlineto identify an efficient method to forecast the future cash demand. The newlinex newlineoptimized cash management model using particle swarm optimization was newlineimplemented for two different data set. The optimized results were compared newlinewith statistical models to prove the accuracy of proposed PSO based cash newlinemanagement with the best accuracy of 70%and 91% for both short term and newlinelong term data. The neural network based cash forecasting model was newlinedesigned to improve the accuracy of PSO based coefficients by intro