Abstract: Based on the fact that personnel competence is the key factor affecting work performance, the renewable resources in resource constrained project scheduling problem are transformed into human resources including competence difference by a series of scientific and reasonable methods or means, and then a model of human resource constrained project scheduling problem with emphasis on competence difference is constructed. The most outstanding advantage of this model is that it can select the objective and reasonable evaluation index of personnel's competence to provide a rigorous and scientific relationship, which transforms the double objective minimization problem of complex multi project total duration and total cost into a single objective maximization problem of comprehensive index, establishes a mathematical optimization model, and uses genetic algorithm to solve it. Compared with the traditional multi-mode model, the model based on the difference of the competency level of personnel is obviously better, and its advantage is that the optimal time limit is shorter and the optimal cost is lower. The mathematical optimization model considering the difference of competence is more in line with the practice of R & D project group management. At the same time, genetic algorithm is not only efficient in solution, but also easier to get objective and accurate results.

Key words: competence, R&, D program, project scheduling, genetic algorithm, serial schedule generation scheme

摘要： 基于人员胜任力是影响工作绩效的关键因素,将资源受限项目调度问题中的可更新资源通过一系列科学合理的方法或者手段转变为存在胜任力差异的人力资源,由此构建起一个强调胜任力差异的人力资源约束项目调度问题模型,此模型最突出的优势在于选取了能够客观合理评估人员胜任力的指标,提供了严谨科学的关系式,将复杂的多项目总工期与总成本的双目标最小化问题转换为综合指标单目标最大化问题,建立数学优化模型,采用遗传算法求解。通过算例研究证实,相较于传统多模式模型,基于人员胜任力水平差异的模型明显更胜一筹,其优势集中表现为最优工期更短、最优成本更低。考虑了胜任力差异的数学优化模型更符合研发项目群管理实践,同时遗传算法在求解方面不仅效率高,并且更容易获得客观准确的结果。

关键词: 胜任力, 研发项目群, 项目调度, 遗传算法, 串行调度生成机制