(708b) Multi-Mode Resource Constrained Project Scheduling for Pharmaceutical R&D

Wang, H., Carnegie Mellon University
Viswanath, S., Eli Lilly and Company
Guntz, S., Eli Lilly and Company
Dieringer, J., Eli Lilly and Company
Vaidyaraman, S., Eli Lilly and Company
Gounaris, C. E., Carnegie Mellon University
Major pharmaceutical companies and their R&D management constantly face a complicated scheduling and resource allocation problem, given the large portfolio of molecules for which they have to pursue simultaneous development. Research and development for each molecule involves multiple inter-dependent activities that are often constrained with tight deliverable due dates. Furthermore, limited resources, such as human resources and allocated budgets among others, are shared across the portfolio. In this context, it is of interest to identify R&D schedules that are optimal against certain performance metrics, including overall net present cost of activities and various other business driven makespan-type objectives. Evidently, the presence of many complicating and sometimes conflicting constraints gives rise to rather challenging scheduling instances.

The multi-mode resource constrained project scheduling problem (MMRCPSP), a very well-studied in the literature archetypal optimization problem, has a broad application in planning and process level enterprise decision problems [1]. The MMRCPSP was first introduced in the context of optimizing pharmaceutical research projects by Kolisch et al. [2], where the authors proposed two heuristics. Other works later implemented MILP models based on MMRCPSP within a heuristic simulation-optimization framework for addressing R&D pipeline management [3,4]. While some studies viewed the single product development as a resource constrained problem [5], other studies on portfolio-wide optimization are using stochastic programming methods [6]. For example, Colvin et al. [7] addressed the similarities shared between R&D planning and RCPSP, while they used a stochastic programming framework to account for the success or not of various clinical trials.

Having appropriately codified the R&D activities that must be carried out before any new pharmaceutical product can undergo its associated clinical trials and planned product launch, we propose here a new mixed integer linear optimization model to solve the portfolio-wide scheduling and resourcing problem for pharmaceutical R&D. The model augments the archetypal MMRCPSP to account for important realities faced by practitioners in this space, including among others inter-activity overlaps beyond standard precedence relationships, resources of mixed (both intensive and extensive) type, as well as optional activities. We demonstrate the tractability of this model and its ability to address portfolio-wide instances, while we deploy a decision support tool for the systematic and largely automated derivation of optimal activity schedules and resource allocations. By utilizing the tool under different input instances, we can also conduct various strategic analyses to assess the system’s ability to cope with sudden changes in portfolio size and/or resource availability.


[1] Varma, Vishal A., et al. "Enterprise-wide modeling & optimization—An overview of emerging research challenges and opportunities." Computers & Chemical Engineering 31.5-6 (2007): 692-711.

[2] Kolisch, Rainer, and Konrad Meyer. "Selection and scheduling of pharmaceutical research projects." Perspectives in modern project scheduling. Springer, Boston, MA, 2006. 321-344.

[3] Varma, Vishal A., et al. "A framework for addressing stochastic and combinatorial aspects of scheduling and resource allocation in pharmaceutical R&D pipelines." Computers & Chemical Engineering 32.4-5 (2008): 1000-1015.

[4] Zapata, Juan Camilo, Vishal A. Varma, and Gintaras V. Reklaitis. "Impact of tactical and operational policies in the selection of a new product portfolio." Computers & Chemical Engineering 32.1-2 (2008): 307-319

[5] Jain, Vipul, and Ignacio E. Grossmann. "Resource-constrained scheduling of tests in new product development." Industrial & Engineering Chemistry Research 38.8 (1999): 3013-3026.

[6] Rogers, Michael J., Anshuman Gupta, and Costas D. Maranas. "Real options based analysis of optimal pharmaceutical research and development portfolios." Industrial & engineering chemistry research 41.25 (2002): 6607-6620.

[7] Colvin, Matthew, and Christos T. Maravelias. "R&D pipeline management: Task interdependencies and risk management." European Journal of Operational Research 215.3 (2011): 616-628.