A heuristic approach for rework based product design project scheduling problem

Product design project scheduling is a complex problem which involves a large number of tasks and information exchanges in a stochastic, resource-constrained project network where iteration and overlap between tasks often take place. The classical determined RCPSP which has been widely studied is based on assumption of fixed duration of tasks and zero-lag finish-start precedence relationships in acyclic AON network. Very little efforts have been made in project scheduling with stochastic duration, especially in cyclic network with rework possibility. This paper presents an approach for rework-based product design project scheduling problem, taking overlap and iteration which are inherent nature in PDP into account. We consider this problem under the assumption that task duration can be modeled by a triangle distribution and can be sampled by simulation techniques. Three priority rules which are widely used in solving deterministic RCPSP are modified by replacing deterministic duration by a dynamic one. Finally computational experiments for this heuristic approach are provided.