טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentPnina Feldman
SubjectConstruction and Evaluation of a Deterministic Model for
Cost Estimation of Manufactured Parts Depending on
the Dynamic Shop-Floor Load
DepartmentDepartment of Industrial Engineering and Management
Supervisor Full Professor Shtub Avraham


Abstract

Cost estimation is a practical problem that has attracted the attention of many researchers. Past research has focused on cost estimation in an unlimited capacity environment. In this research, we present a mixed integer - programming model (MIP) for estimating cost under finite capacity. Because of the high complexity of the problem, we also propose an iterative heuristic model, which combines a tactical level aggregate planning model integrated with a finite loading procedure. The cost estimation using either the optimal model or the heuristic is based on a marginal analysis where the total cost without the new order is subtracted from the total cost with the new order. The proposed heuristic’s efficiency and effectiveness were examined with respect to the optimal solution for small instances of the problem. For those problems both models derived the same objective function values, but the heuristic obtained these solutions faster by several orders of magnitude. For larger problem instances, the heuristic was compared to a partial enumeration solution and to the lower bound obtained from the partial Branch and Bound procedure. For all tested problems, the objective function values obtained by the proposed heuristic were lower than the values obtained by the truncated MIP heuristic and they were not far from the lower bounds values, which may not be feasible. A computational study was performed to examine the benefit of using the proposed heuristic rather than a conventional method. Several factors reflecting different states of shop-floor load were examined using a full factorial design. The results revealed that when the load was high the gain obtained from when using the proposed model was more significant.