| Abstract: |
Resource-constrained project scheduling problem (RCPSP) is normally characterized by makespan length as objective functions and features of resource limitations and permutation condition. The heuristic procedures broadly fall into two categories namely constructive and improvement heuristics. Schedule generation schemes (SGS) are the core of most heuristic solution procedures of RCPSP and MRCPSP. SGS start from scratch and build a feasible schedule. In our study we have resumed that the best heuristics in our computational analysis are metaheuristics based on the activity list representation and the serial SGS. In the current work some new constructive serial scheduling schemes are proposed and others are adapted for multiple resources constrained scheduling process. The performance of these scheduling schemes are evaluated and appeared promising tendencies in scheduling process for single and multiple oriented critical resources(s). R53, RQ, R43, and R26 have achieved the best performance allover the categories under consideration. R43 and ACTIM have achieved the best performance for first category while R52, R53, RQ, MWCA, MTG2, MROT-MACTRES, and MACROS have achieved the best performance for second category. MWROR, MTOR, R25, R31, R41, R49, R52, R53, and R55 have achieved the best performance for third category. The proposed scheduling schemes are integrated in a coded algorithm not only for the evaluation process of the priority orders of the project activities under consideration for n oriented critical resources but also for determining the best schedule for the project under consideration using the available variability of scheduling schemes.Also in this work, a proposed genetic algorithm is presented for scheduling single and multiple resource-constrained projects. The proposed algorithm is tested using fifty test problems and compared with the recommended makespans derived by the proposed fifty-five heuristic rules cited above. The proposed genetic algorithm achieved 70% results better than that rendered by the test benchmark problems while the average deviation per project from heuristics is 0.097% per project for the remainder 30%. Hence the proposed genetic algorithm recommended as a powerful tool for scheduling single and multiple resource constrained projects.5.2 Recommendation for Future WorkIt is hoped that this dissertation will serve as the basis for the analysis and scheduling of single and multiple constrained resource projects. In particular, future research in the following topics can be beneficial:1. Multi mode activity scheduling projects where the project activities can be executed in different modes, involving different durations and resource requirements. This problem is known as MMRCSP and it is NP-hard problem.2. Multi project scheduling problem where the scheduling process is carried out and correlated to allocation process of the shared resources in multi-projects. Scheduling multi-projects with shared resources is quite relevant in practice given that the heuristic solution is the only one possible and the implicit or explicit consideration of each project from multi-project is a relevant factor in scheduling method efficiency.3. Multi mode activity-Multi project scheduling represents also an attractive category of research in the field of project scheduling. In this aspect the heuristics will be of critical concern in solving such category.4. Scheduling projects of multiple resource categories. In project scheduling a common distinction is made between renewable resources, non renewable resources and doubly constrained resources. Renewable resources are available on a period by period basis, that is, the available amount is renewed from period to period. The total resource usage at every time instant is constrained. Non renewable resources are available on a total project basis, with limited consumption availability for the entire project. Doubly constrained resources are constrained per period as well as for the overall project. In this category, project scheduling represents also an attractive category of research and heuristics will be of critical concern in solving such category.5. Other different topics are considered of hot areas of research in project scheduling such as resource constrained project scheduling with multiple objectives, multi mode resource constrained project scheduling with multiple objectives, and project scheduling with stochastic structure and or/ activity interruptions of stochastic nature.
|
|
|