یک مدل بسته بندی چند جعبه ای برای انتخاب پروژه و برنامه ریزی "تحقیق و توسعه"

R&D project selection has been a popular research topic since the 1960s, but the scheduling issue has rarely been considered in such research. This paper records the development and application of a heuristic packing-multiple-boxes (PMB, or multi-knapsacks-model) model, which is based on the packing-single-box (PSB) model. The PMB model can be used for both selecting and scheduling candidate R&D projects. The model was established at the request of a case company that conducts up to 20 large R&D projects simultaneously. The example of using this model at the case company for a five-year planning period is provided. Although the model may not be a optimistic one, it provides one solution to solve this problem.

This paper is based on research applied to a case company, here referred to as Company H. Company H conducts large R&D projects that are characterized by the following three features. First, the projects consume up to tens of millions US dollars. Second, the projects' time spans last from several years to decades with highly variable annual costs. Third, many large projects are conducted simultaneously. At the beginning of every five-year planning period, Company H faces a problem of selecting, as well as scheduling, candidate R&D projects in order to meet the budgetary limitations. At the time this research started, Company H had eight candidate R&D projects for the forthcoming five-year period. Their total and annual costs in this planning period are shown in Table 1, with all the data having been filtered for reasons of confidentiality. Although projects P1, P2, and P3 were started in the previous planning period, such ongoing projects may be suspended or even cancelled if their contributions are lower than new candidate projects after the re-evaluation prior to the forthcoming planning period. As such, they are also regarded as candidate projects at the beginning of every planning period. Projects P4–P8 are new candidates, and P5 and P6 will last more than five years. It should be noted that the procedure for estimation of total costs is omitted in this paper, but readers who are interested in project cost estimation are directed to references including those of Humphreys and English, 1993 and Greer and Nussbaum, 1990.

This article records the idea and development of a packing-multiple-boxes model as applied to the selection and scheduling of R&D projects. The model is based on several simple 0–1 integer programming models. The packing-multiple-boxes model can be used to select R&D projects and, at the same time, to make a schedule of the selected projects under specified financial constraints. The packing-multiple-boxes model greatly improves the traditional ‘packing (single) box model.’ The main contribution of the packing-multiple-boxes model is the incorporation of both the selection and scheduling of R&D projects, which solves the deficiency of earlier approaches which are unable to perform selection and scheduling simultaneously. The separation of R&D project selection from scheduling tasks in previous models was based on the assumption that project selection and scheduling are independent and the projects selected are not interrelated in terms of resources during the scheduling process. However, the problem in the case company of the research presented here reveals that project selection and scheduling should be regarded as being dependent and closely related.