关键词: 拖期无知, 资源受限项目调度, 多模式调度, 广义证据融合规则

Abstract: With the increasing complexity of project implementation and resources occupation, external uncertainties and risks are increasing. A reasonable scheduling of limited resources and activity has become the key to a project scheduling. In order to better guide social production practice, scholars further expand the model of the resources-constrained project scheduling problem (RCPSP) and study related issues. Due to the uncertainty of the project implementation environment, activities are faced with inaccurate time estimation, insufficient resources update and other human factors, which delays the project activities. The tardiness of activities will cause the uncertain risks and costs of the project to rise, and even have a “ripple effect” on the follow-up activities, which will affect the overall stable operation of the project. Therefore, experts are needed to estimate the tardiness probability in the ignorant situation. However, the current research generally assumes that in the project there is an experience to follow when judging whether an activity is delayed, and the experts who judge the tardiness probability are omniscient. However, when a real problem is in an ignorant situation, they will not be able to accurately express the tardiness probability of the project activity. In this context, addressing the challenge of scheduling under such conditions of ignorance becomes crucial for the successful management of projects. In this paper, combined with the generalized combination rule, the ignorance of expert prediction is overcome, and the scheduling method and application of multi-mode resources-constrained projects with ignorant tardiness duration is studied. It provides a new idea for solving the project scheduling problem under uncertain circumstances and provides a basis for the corresponding management practice.
In this paper, a robust optimization model of the multi-mode resources-constrained project scheduling problem with tardiness ignorance is constructed, and the genetic algorithm is used to optimize it. By maximizing the robustness of the project, we can reduce the influence caused by the activity delay and ensure the stability of the project operation. In the process of calculation, it is drawn that the greater the resources of an activity, the greater the importance of the activity in the project and the higher the priority. This correlation is particularly important as it allows for the prioritization of resources based on the criticality of the activities, which in turn can significantly impact the overall project timeline and success. This paper assumes that there is a positive relationship between the importance of an activity and the allocation of renewable and non-renewable resources it needs. Through the robust optimization of project tardiness, this paper explores how to carry out and reasonably determine the duration-resources model of each project activity under the constraints of activity priority, project duration, renewable and non-renewable resources.
In order to verify the validity of the model, this paper constructs an ecological security supervision network of marine ranching. As a key measure to develop marine economy and ecological fishery in China, marine ranching plays a significant role in conserving fishery resources, coping with environmental damage and improving ecological benefits, but in recent years, ecological problems of marine ranching have arisen frequently. Due to the nascent stage of marine ranching and scarcity of historical data, optimizing the robustness of the ecological security supervision network is not only a theoretical challenge but also a practical necessity. Due to a late start of marine ranching construction and lack of historical experience and achievements, it is of practical significance to optimize the robustness of the marine ranching ecological security supervision network with delay and ignorance. In this paper, the work breakdown structure method is used to divide the marine ranching ecological supervision project into four stages: the design stage, the marine ranching ecological security monitoring and evaluation project, the marine ranching ecological security early warning project and the marine ranching ecological emergency decision-making project. The final optimization results verify the effectiveness of the robust optimization model of tardiness-ignorant MRCPSP, and provide theoretical guidance for the development of inexperienced projects.

Key words: ignorant tardiness duration, resources-constrained project scheduling, multi-mode scheduling, generalized combination rule