摘要
项目调度计划中不利的资源波动往往导致项目施工现场执行中的资源效率低下和成本上升。提出一个创新的资源均衡度量指标:资源波动成本,从而可以直接度量并最小化资源波动对施工生产力和成本造成的负面影响。该指标在出现不利的资源波动情况下,分别考虑采取允许资源闲置和允许解雇后再雇用两种不同的资源使用策略。建立一个以资源波动成本最小化为目标的资源均衡优化模型,比较不同策略下的资源闲置成本和解雇再雇用资源成本,确定使总资源波动成本最小的资源使用策略;结合问题的实际特点,设计求解模型的遗传算法;最后,用一个实例说明模型的合理性和算法的有效性;算例对比分析结果表明所提出的资源均衡度量指标优于现有的度量指标,不仅可以最大限度地消除不利的资源波动和资源闲置时间,而且可以根据实际情况确定相应的资源使用策略。
The occurrance of resource fiuctuation in project scheduling often leads to the inefficient utilization of resource and rise in cost on cunstruction sites. In order to measure the resource leveling, many metrics have been put forward, such as, minimum variance, minimum moment, minimum sum of squares, minimum absolute difference etc. However, all the aforementioned metrics imply an assumption： it is the peak of resource profiles that determines the lowest amount of hired resource, which means there might be idle resource at the valley of resource profiles. Nevertheless, on real construction site, it might be considered the release and rehire strategy, which is to temporarily release resource during low demand periods and rehire again during later high demand periods. Of course, the adoption of this strategy depends on the comparison between the unit cost of idle resources and that of resource fluctuation. Therefore, the objective of this paper is to present an innovative resource leveling metric which can reflect resource fluctuation cost directly, and generate corresponding optimal scheduling and resource utilization strategy. In the first part, this paper describes and elaborates the meaning and the computational rule of resource fiuetuation cost. Resource fluctuation cost equals to the minimal value of the two resource fluctuation costs when either adopting the release and rehire strategy or allowing resource idle strategy. In the second part, we build a resource leveling optimization model, to minimize the resource fluctuation cost subject to precedence and floats constraints. Later, in the third part, an improved genetic algorithm is proposed to solve this problem after considering the actual characteristics of resource leveling problem. The chromosome of the new algorithm is encoded by the priority list and actual shifted days list of the non-critical activities. This encoding mode can not only generate the start time of the non-critical activities randomly, but also meet the requirements of precedence co
出处
《管理工程学报》
CSSCI
北大核心
2015年第2期167-174,共8页
Journal of Industrial Engineering and Engineering Management
基金
国家自然科学基金资助项目(70871088
71272146)
关键词
资源均衡优化
资源波动成本
解雇再雇用策略
资源闲置策略
遗传算法
resource leveling optimization
resource fluctuation cost
release and rehire strategy
resource idle strategy
genetic algorithm
