An examination of membership function performance in predicting the ideal order quantity for perishable items

Fuzzy logic is an effort to imitate how people make decisions and solve problems. Fuzzy logic is based on human thinking since it relies on values that are not absolute truth or falsity, but rather on values that are vague or ambiguous. Numerous studies have demonstrated that linguistic information that is ambiguous or unclear may be processed by fuzzy logic. Numerous domains, such as medical diagnostics, control theory, management, artificial intelligence operations, and more have used the fuzzy logic method. Selecting the right membership function is seen to be a difficult undertaking since it influences the fuzzy inference system’s performance and the solution to the specific problem at hand. In our present work, we have fuzzified the data into fuzzy sets for both input and output, using just the triangle membership function. In the second case, we utilize the Trapezoidal membership function. In a similar vein, we study three kinds of membership functions (Gaussian, Trapezoidal, and Triangular) for input and output variable fuzzification. Next, we calculate each inference system’s error percentage between observed and forecasted values, comparing the outcomes to ascertain each inference system’s accuracy percentage. To illustrate the performance evaluation of these membership functions, we provide a graphical representation.