INFLUENCE OF DATA QUANTITY ON ACCURACY OF PREDICTIONS IN MODELING TOOL LIFE BY THE USE OF GENETIC ALGORITHMS

It is widely known that genetic algorithms can be used in search space and modeling problems. In this paper theirs ability to model a function while varying the amount of input data is tested. Function which is used for this research is a tool life function. This concept is chosen because by being able to predict tool life, workshops can optimize their production rate – expenses ratio. Also they would gain profit by minimizing number of experiments necessary for acquiring enough input data in process of modeling tool life function. Tool life by its nature is a multiple factor dependent problem. By using four factors, to acquire adequate tool life function, vivid complexity is simulated while acceptable duration of computational time is maintained. As a result almost clear threshold, of data quantity inputted in optimization model to gain acceptable results in means of output function accuracy, is noticed.