干扰管理视角下大型复杂工程项目多目标鲁棒调度优化

During the implementation of large and complex construction projects (LC-CP), the realized progress of projects over time very easily deviates from the baseline schedule, or may be out of control because the concurrent appearance of various uncertain disruption events, especially for complicated interrelated relationships existing among these events. Therefore, it is very important to enhance the evolution stability along a LC-CP execution to gain expected multi-objective. Therefore, in the new perspective of disruption management (DM), our scientific program investigates on robust scheduling and multi-objective optimization of LC-C under uncertain environments. Firstly, the impacts of various disruption events on project implementations are coupled and interactive, and disruption events often arisen at the multiple stages of project execution process. We explore the basic discipline and mathematical expression for these disturbance effects. Then, the robust multi-objective scheduling optimization models of project DM are constructed based on the uncertain mixed integer programming, the constraint programming, and so on, and the solution and quality robust performances are simultaneously contained in the modes. Accordingly, some intervention options are designed for the models, such as alternatives for activity’s executing modes, resource substitution, etc. Next, several quickly local rescheduling algorithms are developed to solve the models, which integrate the discrete simulation and the heuristic intelligent optimization techniques. Some on-line recovery options can be obtained by the algorithms. Finally, the research results will be applied in the real context of project management, and forming a new system of project scheduling theory and method. On the one hand, the efforts in this research can provide a decision support tool for accurately and effectively controlling the evolution of a LC-CP; on the other hand, we theoretically extend the model of DM in order to improve its application for more complicated system under the disturbance and recovery status.

大型复杂工程项目实施中，各种干扰事件交织在复杂的关联关系中，极易导致进度偏离计划甚至失控，提高其执行过程的稳定性对预期多种目标的实现很关键。本项目以干扰管理为切入点，研究不确定环境下大型复杂工程项目多目标鲁棒调度优化问题。探索项目在执行中的多个阶段出现干扰事件、多种干扰事件之间交互耦合对项目进展状态产生扰动的作用规律和数学表达；在此基础上，利用不确定性混合整数规划和约束规划等理论，构建同时包含解鲁棒性和质量鲁棒性的多目标项目干扰调度优化模型，设计活动的执行模式替代、资源替代等多种干预方案；开发将离散仿真和启发式智能优化技术相结合的快速局部重调度算法，获得多个阶段的在线扰动恢复策略；最后将研究结果应用于实际项目管理情境，形成一种新的项目进度管理理论与方法。研究成果为高效、准确地实施大型复杂工程项目的进度控制提供了决策工具，并从理论上拓展了干扰管理模型以使其适用于对复杂系统的扰动恢复决策。

大型复杂工程项目实施中，各种不确定性干扰事件导致实际进度偏离计划甚至失控，提高其执行过程的稳定性对实现多种预期的目标很关键。识别干扰事件对进度执行的影响程度、量化其影响进而构建多个相关问题的混合整数规模模型。首先，用时差效用函数的新指标度量调度解的鲁棒性，并同时考虑鲁棒性和项目工期，构建问题的整数规划模型并设计求解算法；从提高调度过程的稳定性角度，建模并求解随机活动工期的资源约束型鲁棒性项目调度问题；从前摄性和反应性集成角度，以最大化期望净现值为目标研究随机活动工期下的项目调度问题。其次，针对关键链在理论和实践两方面的瓶颈问题（“二次资源困境”），开发出一种消除二次资源冲突的启发式局部重调度协调策略，并设计基于策略的鲁棒性度量指标；建立以鲁棒性为目标的关键链调度新模型并求解，形成的鲁棒性关键链调度理论创新了传统的关键链方法。再次，从多项目角度研究大型复杂工程项目：针对分布式资源受限的多项目调度问题，开发了基于关键链改进搜索的遗传算法求解；进一步在问题中考虑全局资源的转移时间，区分静态和动态的项目到达场景，分别构建两种场景下问题的整数规划模型并设计有效的求解算法。再其次，对以地铁修建为典型代表的大型工程项目进度管理问题，构建了时间-空间冲突下的空间资源约束型项目调度模型，进而设计了一种适合解决空间资源配置的启发式遗传算法求解模型。最后，对六个典型工程项目开展实地调研及案例研究，取得了较好的效果并完善了课题的理论研究成果。在复杂多变的环境下，研究成果为高效、准确地实施大型复杂工程项目进度控制提供决策支持，具有重要的现实意义和广泛的应用前景。