Scheduling of design projects with uncertain number of iterations

A short product design cycle is critical to the success of companies in the era of time-based competition. The underlying design activities, however, are often interlinked and quite uncertain. For example, some activities may have to be iterated several times to meet the design criteria. Furthermore, time-critical projects suffer the risk of failure if they cannot meet established target dates. Generating good and robust schedules is thus critical, especially under the concurrent engineering paradigm where the delay of a single task may have a domino effect on subsequent tasks and on other projects sharing designers and/or resources. This paper studies the scheduling of design projects with uncertain number of iterations while managing design risks. A ‘separable’ problem formulation that balances modeling accuracy and computation complexity is created with the goal to minimize project tardiness and risk penalties. An optimization-based methodology that combines Lagrangian relaxation, stochastic dynamic programming, and ‘ordinal optimization’ is developed. Numerical results supported by simulation demonstrate that near optimal solutions are obtained, and uncertainties are effectively managed for problems of practical sizes.