the 12 DRCMPSP instance sets (totaling 60 instances) in the Multi-Project Scheduling Problem Library (MPSPLIB)

In response to issues such as severe resource conflicts and prominent project delay risks arising from the leapfrog development of the aerospace industry, and considering the real-world scenarios where new projects arrive unpredictably and ongoing project plans change unexpectedly during multi-project execution, this paper proposes the problem of dynamic distributed multi-project resource configuration for aerospace missions. Combining the characteristics of aerospace project tasks and the practical application scenarios of multi-project management, project weights are introduced. A dynamic multi-project resource configuration management process based on Multi-Agent Systems (MAS) and the Critical Chain Project Management (CCPM) method is proposed. A three-stage iterative model is established with the objectives of optimal change scope, shortest single-project duration, and minimal multi-project weighted tardiness. This model comprises reconstruction scope definition, local single-project scheduling, and global multi-project decision-making. A genetic algorithm incorporating optimal inheritance and project weights is designed to solve the model. Based on the DRCMPSP instances from the MPSPLIB benchmark library, a dynamic multi-project resource configuration test set is constructed through parameter configuration, and large-scale numerical experiments are conducted. The results show that the proposed dynamic multi-project resource configuration model and its solution algorithm can effectively mitigate the negative impacts caused by resource conflicts and project delays in aerospace multi-project management. It demonstrates good adaptability to different problem scales and resource utilization coefficients, proving particularly suitable for situations with severe resource conflicts, which aligns with current and future development trends.针对航天工业跨越式发展过程中出现的资源冲突严重、项目延期风险突出等问题，结合多项目执行过程中新项目随机到访、在研项目计划突发变更的现实场景，本文针对航天任务场景下的动态分布式多项目资源配置问题展开研究。结合航天项目任务的特性与多项目管理的实际应用场景，引入项目权重，提出了基于多智能体系统（Multi-Agent Systems, MAS）与关键链项目管理（Critical Chain Project Management, CCPM）方法的动态多项目资源配置管理流程。以变更范围最优、单项目工期最短、多项目加权拖期最小为目标，构建了三阶段迭代模型，该模型涵盖范围重构定义、局部单项目调度以及全局多项目决策三个环节。设计了融合最优继承策略与项目权重的遗传算法对该模型进行求解。基于MPSPLIB基准库中的DRCMPSP实例，通过参数配置构建动态多项目资源配置测试集，并开展大规模数值实验。实验结果表明，所提出的动态多项目资源配置模型及其求解算法可有效缓解航天多项目管理中资源冲突与项目延期带来的负面影响，对不同问题规模与资源利用系数均具备良好的适应性，尤其适用于资源冲突严重的场景，契合当前及未来的发展趋势。