Abstract: The original production plan is a production plan that minimizes the cost to meet the needs of forecasted product types and quantities. However, the actual demand often does not match the forecast, and the original production plan needs to be dynamically adjusted. Based on this, this paper, by utilizing the inverse optimization methodology, transforms the resource reconfiguration problem for production plan adjustment into a problem of inversion of resource parameters, which is a known practical optimal production plan, and the reverse solution makes the plan the optimal solution of the original production planning model under the new parameters. Taking the changes in market demand as the driving force, with the goal of adjusting the minimum amplitude of the production time parameters of various types of products of each production line, the adjustment cost is minimized from the two dimensions of resource change quantity and change item, the production resources are reallocated, and the resource adjustment cost is minimized while meeting new market demands. The inverse optimization model is a non-convex nonlinear problem. The three-stage method of “relaxation constraint-median constraint-tight relaxation constraint” determines the production capacity constraint of the production line, which can enhance the accuracy of the optimal solution and improve the solution efficiency. In order to minimize resource adjustment, the   norm is introduced to reduce the number of resource changes and further reduce the cost of production changes. The global equipment efficiency OEE is used to measure the production efficiency of the operation link, making the production resource allocation process more global and systematic. This paper introduces inverse optimization into the adjustment of production resources driven by goal factors and market changes demand, which helps to promote supply and demand matching, stabilizes production, and reduces the cost of enterprise production adjustment.

Key words: production planning adjustment；resource reconfiguration；inverse optimization；  , norm regularization；three-stage capacity verification

摘要： 原生产计划是满足预测产品种类及数量需求的成本最小化的生产计划。然而，实际需求常常与预测不符，当获取实际需求数据后，需要对原生产计划进行动态调整。基于此，以逆优化为方法论，将面向生产计划调整的资源重配置问题，转化为一个已知实践最优生产计划，反向求解让该实践方案成为在调整后的参数下的原生产计划模型最优解的反演问题。该反演问题是以市场需求变动为动因，以不同生产线上不同类型产品的生产时间为调整参数，通过对该参数最小化调整的同时达到市场实际需求。首先从资源种类变动项目和每项资源变动数量两个维度刻画调整成本，同时引入 范数正则化方法降低资源变动项数，降低资源变动成本；其次，通过采用全局设备效率OEE测量作业环节生产效率，使得生产资源配置过程更具有全局性与系统性；最后，鉴于本逆优化模型是非凸非线性问题，为了提高求解效率，提出“松弛约束-中位数约束-紧松弛约束”三阶段方法确定生产线生产能力约束，该方法能够在增强最优解准确性的同时提高求解效率。将逆优化引入由果索因、变动需求驱动的生产资源最小化调整问题，有助于促进供需匹配，稳定生产，降低企业生产调整成本。

关键词: 生产计划调整, 资源重配置, 逆优化, 范数正则化, 三阶段产能核定