|M.Sc Student||Ben-Shabat Shlomo|
|Subject||A Simulator for Training in System Engineering|
|Department||Department of Industrial Engineering and Management||Supervisor||Professor Avraham Shtub|
|Full Thesis text - in Hebrew|
In light of the increasing demand for engineers with system engineering (SE) skills, many training programs dedicated to SE have been opened in academia and industry. Unfortunately, due to shortage of simulation tools for learning and practicing engineering systems, instructors have to teach big majority of the learning material through frontal lessons without using any appropriate training tool. In order to meet this need a SE training tool was developed as part of this research. The research’s purpose was to investigate whether the simulator can demonstrate the effects of different system design strategies on project’s results under different uncertainty conditions.
The study was carried out by a controlled experiment using two system design projects which had been uploaded to the simulator. The experimenters were 32 graduate students in the SE ME program at the Technion. Using the simulation results we examined the relationship between the objective functions which were defined to the students and the project results in terms of cost and system performance. The design of the experiment was based on two design strategies "Design to Cost" (DTC) and Cost Benefit (CB).
By comparing the results of the two methods we found out that the students of the CB group that were measured by the Benefit and cost, had completed their simulation runs with lower cost and lower benefit than DTC group.
The results illustrate the effect of the uncertainty level on how the students have managed the projects according to the defined objective functions, and accordingly on the differences in the outcomes between the groups.
As a result, the difference between the system's performances of the groups is smaller than the difference in performance in a scenario with a low level of uncertainty . However, despite the differences in performance, differences in cost between the groups are very similar in both scenarios because the cost that resulted from treatment of risks was expended at the expense of performance.
Efficiency frontier for both scenarios includes results of the two strategies. Comparing the average distance of each strategy's results from the efficiency frontier shows that there is no advantage of either strategy in term of the chances to get closer to the efficiency frontier. However, choosing the strategy affects on the area on the efficiency frontier where the result will be placed.