بهینهسازی در زمانبندی پروژه تصادفی
Abstract
1-Introduction
2-Literature Review
3-Theories and Methodology
4-Applications of Response Surface Methodology
5-Conclusion
References
Abstract
In the field of project scheduling, the application of activity floats is often the subject of researches that seek to maximize the net present value (NPV) of project networks. However, NPV-based scheduling in stochastic projects may result in a conflict with the primary objective of scheduling, i.e. minimizing the makespan. The main objective of this study is to improve the financial gain while respecting the makespan (i.e. applying safe floats) in stochastic projects. Next, instead of the traditional use of float which prescribes a fixed delay in the start of activity in any condition, we use release dates as the by-product of above-mentioned method to reduce the variation in time and NPV of the project with less computational effort. The project models are simulated in SAS Simulation Studio. Response Surface Methodology (RSM) is used to design the experiments, interpret the results and predict the solution. The focus is on the temporal analysis of stochastic networks, and no resource constraint is considered.
Introduction
Delaying the start of activities postpones the time a cost incurs, and consequently leads to higher net present value of the project which indicates the present value of future money. Other benefits of such delays (e.g. the reduction in holding expenses) may also apply depending on the characteristics of the real life situation. As we will see in the third example, the true application of delays reduces the variation in the schedule. Managing a project usually needs coordination among activities, so variation is not very welcomed. Jorgensen and Wallace (2000) truly discuss that, according to Jensen’s inequality, planning methods based on the use of average numbers may lead to underestimation of project costs and durations. So, delaying an activity by a float calculated based on the Pert/CPM procedure may lead to postponement of the project completion. The first two examples, borrowed from the pioneering work of Buss and Rosenblatt (1997) on the subject, will show that NPVbased scheduling may also postpone the project completion. Another issue about NPV-based scheduling is that any activity is a potential candidate for optimizing the objective, even if the activity is critical.