Application of neural networks to heuristic scheduling algorithms

As its name implies, this thesis presents application of artificial neural networks to six heuristic scheduling algorithms with the makespan objective fimction.(CDS algorithm, Neh algorithm, Koulamas' algorithm, DEMIR's Frequency algorithm, DEMIR's Point algorithm, Aksoy's algorithm) In this thesis, according to the Axle Housing Workshop Global Operation Recipe in Ege Endüstri Factory, the best sequence for 5 jobs processed on 43 serial machines are found by these six heuristic methods. For every sequence found by six different heuristics, the completion time of each job on each machine, job waiting times and machine idle times are computed. In order to develop a neural network, the Backpack Neural Network System Version 4.0(by Z Solutions) is used and some necessary steps are followed. The data is read into the system. Job and machine numbers, processing times of each job on each machine, completion times of each job on each machine, job waiting times, machine idle times are considered as our data. Then desired preprocessing on the data is performed using the fuzzy component and the special component. A 1 of N transformation on jobs and machines are created, fuzzy variables to represent completion times, job waiting and machine idle times are determined. In this study, triangular fuzzy numbers are used to represent the fuzzy completion times. Job waiting and machine idle times are considered as triangular or trapezoidal fuzzy numbers. The data is split into "train", "test" and "validate" datasets. The output variable, completion time(earlyf) is chosen, the number of groups(stratas) are entered to stratify the datasets. For each heuristic with different number of strata(between 3 and 6), the neural network is trained using the Train component. Then the network is evaluated using the Validate dataset and it is seen how well the model works, using DOKÜMANTASYON MEUQI&the Apply component. After satisfied with the performance of the model, the actual dataset belonging to the completion times of each job on each machine in April 2001 are used to obtain the predictions for the next production period. The predictions of the completion times, fit statistics, graphs of fit are obtained from the model showing the comparison between actual and predicted output values.