On the Development of Theory-Informed Operationalised Definitions of Demand Patterns

不同的库存单位(SKU)与不同的需求模式相关联，而需求模式又需要不同的预测和库存控制方法。因此，我们需要对各种SKU进行分类，并为每个特定类别应用最合适的方法。我们执行这项任务的方式显然对库存和客户满意度具有巨大的影响，因此，相关规则构成了每个库存管理系统的关键要素。为了处理这个问题，人们倾向于相当武断地对需求模式进行分类(使用基于经验、效果良好的规则)，然后选择在每一类中表现最好的预测和库存控制方法。然而，选择最合适的预测和库存控制方法正是进行任何分类工作的目的。因此，更合乎逻辑的做法是，首先比较不同的估算程序和库存控制模型，以确定它们表现优异的区域，然后根据结果对需求模式进行分类，而不是反过来。像这样的程序显然有望提供更好的结果。在与约翰·博伊兰和约翰·克罗斯顿一起进行的研究中，我们开发了一个理论上连贯的分类方案，沿着上面讨论的路线，仅用于预测目的。然而，库存控制问题没有得到解决，这就是我在这项拟议的研究中想要做的。多年来，这一研究领域引起的学术界关注非常有限。造成这种情况的一个原因很可能是相关的复杂性相当大。也就是说，预测和库存控制必须被视为相互关联的功能(就像它们在实践中一样)，而不是更广泛解决方案的独立模块，这显然增加了问题的理论复杂性。虽然对预测和库存控制之间的相互作用已经做了一些早期的工作，但仍然需要一个理论上一致的方法，这是第一个提供这种方法的建议。本研究的主要目的是为预测和库存控制目的产生理论上合理的需求分类规则。要进行这样的项目，来自行业从业者的投入是非常重要的。在这方面，两家公司被选为项目合作伙伴。这种合作还将确保提供本研究所需的经验数据。我的哲学立场是实证主义的，从这个意义上说，寻求开发普遍适用的分类解决方案。然而，由于问题的复杂性，采用的研究策略不能纯粹是演绎的。理论和数据之间的迭代程序将被引入，这种方法将确保确定所有重要的因素。总之，拟议的研究处理了一个从理论和实践角度都值得研究的问题。在需求分类领域开展的工作很少，从研究到目前为止，还不清楚经理们应该如何对需求模式进行分类，以便进行预测和库存控制。这一问题的重要性已在多个场合得到报道，有关文献商定的是迫切需要进一步增进这一领域的知识，并从经验上评估相关问题。因此，拟议的研究是一个非常及时的项目。这一项目的成果将在所有预测和库存控制软件制造商中得到应用。事实上，在任何工业环境中，只要有内部开发或购买的需求分类计算机化解决方案，以促进库存管理，这也是一种自然的应用。

Different Stock Keeping Units (SKUs) are associated with different demand patterns, which in turn require different methods for forecasting and stock control. Consequently, we need to categorise the various SKUs and apply the most appropriate methods for each particular category. The way we are going to perform this task has obviously tremendous implications in terms of stock and customer satisfaction and, as such, the relevant rules constitute a vital element of every inventory management system. To deal with this problem people tend to classify, rather arbitrarily, the demand patterns (using rules that, based on experience, work well) and then select the forecasting and stock control methods that perform best in each category. Nevertheless, the choice of the most appropriate forecasting and inventory control methods is the very purpose of conducting any categorisation exercise. Therefore, it is more logical to first compare alternative estimation procedures and stock control models for the purpose of identifying their regions of superior performance and then, based on the results, categorise the demand patterns, rather than working the other way around. A procedure like this one is obviously expected to offer better results. In research conducted with John Boylan and John Croston we developed a theoretically coherent categorisation scheme, along the lines discussed above, for forecasting purposes only. However, stock control issues were not addressed and this is what I would like to do in this proposed research. This research area has attracted very limited academic attention over the years. A reason for that may well be the considerable associated complexity. That is, forecasting and stock control have to be viewed as interrelated functions (as they are in practice) rather than stand-alone modules of a wider solution, and this obviously increases the theoretical complexity of the problem. Although some early work has been done on the interaction between forecasting and stock control, a theoretically coherent approach is still required and this is the first proposal to provide it.The main objective of this research is to produce theoretically sound demand categorisation rules for both forecasting and stock control purposes. To conduct such a project, the input from industry practitioners is very important. In this regard, two companies have been selected as project partners. This collaboration will also ensure that the empirical data required for the purposes of this research becomes available. My philosophical stand-point is positivistic in the sense that universally applicable categorisation solutions are sought to be developed. However, due to the complexity of the problem, the research strategy employed cannot be purely deductive. An iterative procedure between theory and data is to be introduced and such an approach will ensure that all important factors are identified.In summary, the proposed research deals with an issue that is worth investigating from both a theoretical and practitioner perspective. Very little work has been conducted in the area of demand categorisation and, from the research to date, it is not clear how managers should classify demand patterns for forecasting and stock control purposes. The importance of this issue has been reported on numerous occasions, and what is agreed upon in the relevant literature is the immediate need to further advance knowledge in this area and empirically assess the relevant issues. The proposed research therefore constitutes a very timely project. The results of such a project will find application in all forecasting and stock control software package manufacturers. Indeed, there is also a natural application to any industrial setting where an in-house developed or bought-in demand classification computerised solution is in place to facilitate inventory management.