Distributed simulation of resource constrained project scheduling

Algorithms for resource constrained project scheduling are available for implementation on centralized computer systems. Multiple runs to study the effect of changes in exogenous variables have to be simulated in a serial order on a centralized system. However, inherent parallelism in simulation can be benefically exploited on a distributed computer system to drastically reduce simulation time. Distributed simulation on loosely coupled systems involves problem partitioning into components, allocation of components to computing systems, communication and coordination between components. Strategies for distributed simulation of time analysis of project network have been reported in the authors’ previous work. However, no previous work is reported on the distributed simulation of resource and cost analysis. In this paper, the authors discuss how resource constrained project scheduling can be simulated on a distributed computer system. The process involves partitioning of the project network into independent sub-project networks, and allocating sub-projects to independent computing systems connected through communication channels without shared variables. The total resource availability of the project is maintained at a central computer system. The central system receives activity time vector from individual sub-projects as and when an activity becomes eligible for scheduling and returns the activity schedule vector after the activity has been scheduled. The distributed algorithm proposed has been illustrated by an example.