Taahhütler Altında Proje Planlaması için Dinamik Karar Verme Modeli Yaklaşımı

Abstract

Many companies operating in different fields carry out various projects for the contracts they. Companies make some commitments with these contracts. If these commitments are not fulfilled, penalties and sanctions occur. In order to fulfill the conditions required by the contract, projects must be dynamically planned and managed with limited time, budget and resources. Time commitment has been considered within the scope of the study and Markov decision process model and heuristic approach are proposed. The minimum and maximum workloads of the activities are determined close to reality by utilizing the knowledge and experience of the employees in the project and historical data. The longest time that the project can be completed is calculated by the critical path method by using the predecessor and successor relationship and the calculated maximum durations of the activities. The model optimizes the average longest time for the project by using minimization function and finds the average best time through backward formulation. In a project having heavy workloads, the presence of many serial and parallel activities, the use of too many resources, and a wide range of transition probabilities make it difficult to plan the project dynamically. The progress of the activities and their planned durations were compared with a certain algorithm and a decision rule was developed in which the action to be taken was decided. By using the decision rule, the model is differentiated and a heuristic model is created. Due to the time commitment, the time promised in the contract at the beginning of the project was taken into account.By using the heuristic model, realizations closer to the time promised at the beginning of the project were achieved. In addition, since the action was determined in each period, the computational effort was reduced. The created models were solved for different project networks. According to the activity links and sizes, suitable models for the projects and the parameters to be used have been determined. By comparing the results of the Markov decision model and the heuristic model, the constraint values and parameters in the models can be updated. In order to reduce the processing load, it has been proposed to divide the projects into work packages.