Impact of Asset Utilization on Replacement Decisions

Hartman 9713690 该赠款为优化工具的开发提供资金，以确定随机操作环境下一组资产的替换决策。 传统模型在离散（每年）的时间间隔内提供更换决策，基于预期的运营、维护和抢救成本，而这些成本又取决于运营（日常）决策所确定的资产利用率水平。然而，由于与运营相关的不确定性，例如需求或预算水平的波动，这些预期的利用率水平在实践中可能无法实现，从而使更换计划无效。 该研究将战略（更换）和运营（利用）决策联系起来，使更换计划成为资产利用的函数。 这个问题将在运输行业的背景下进行研究，赞助商包括一家铁路公司和一家卡车租赁公司。 研究了双端排队模型来近似运输系统的运行并确定利用率水平。 该模型内置于非线性优化模型中，该模型确定最低成本的购买和销售决策。 大规模非线性程序的研究求解过程基于增强拉格朗日技术。 这项研究在工业和军事领域有许多潜在的应用，因为这些组织通常依赖于在不确定的环境中受到不同程度的利用的资产组或舰队。 该分析将合并两个传统上独立的决策，即静态替换问题和动态运营问题，以便战略资本预算决策基于实际运营的衡量。 由于收购决策取决于资产利用率，该模型将改进传统上依赖静态数据的资产置换决策过程。 这项研究还将有助于改进大规模非线性优化模型的求解技术，并提供对双端队列的更多研究，尽管双端队列有众多应用，但在研究中并不突出。

Hartman 9713690 This grant provides funding for the development of an optimization tool to determine replacement decisions for a group of assets subject to a random operational environment. Traditional models, which provide replacement decisions over discrete (annual) intervals, are based on expected operating, maintenance and salvage costs which are dependent on asset utilization levels determined by operational (daily) decisions. However, due to uncertainty associated with operations, such as fluctuating demand or budget levels, these expected utilization levels may not be realized in practice, thus invalidating the replacement schedule. The research will link strategic (replacement) and operational (utilization) decisions such that replacement schedules are functions of asset utilization. This problem will be examined in the context of the transportation industry, with sponsors including a railroad and a truck leasing firm. A double-ended queueing model is researched to approximate the operations of the transportation system and determine utilization levels. This model is built into a non-linear optimization model which determines minimum cost purchase and sale decisions. The research solution procedure for the large scale non-linear program is based on augmented Lagrangian techniques. There are numerous potential applications of this research in industry and the military as these organizations generally rely on groups or fleets of assets which are subject to varying levels of utilization in an uncertain environment. The analysis will merge two traditionally separate decisions, the static replacement problem and the dynamic operational problem, such that strategic capital budgeting decisions are based on a measurement of actual operations. As acquisition decisions are dependent on asset utilization, the model will improve the asset replacement decision process which traditionally relies on static data. This research will also contribute to improved solution techniques for large scale non-linear optimization models and provide more investigation into double-ended queues which despite their numerous applications are not prominent in research.