Solving the Resource Constrained Project Scheduling Problem to Minimize the Financial Failure Risk

In practice, a project usually involves cash inand outflows associated with each activity. This paper aims to minimize the payment failure risk during the project execution for the resource-constrained project scheduling problem (RCPSP). In such models, the money-time value, which is the product of the net cash in-flow and the time length from the completion time of each activity to the project deadline, provides a financial evaluation of project cash availability. The cash availability of a project schedule is defined as the sum of these money-time values associated with all activities, which is mathematically equivalent to the minimization objective of total weighted completion time. This paper presents four memetic algorithms (MAs) which differ in the construction of initial population and restart strategy, and a double variable neighborhood search algorithm for solving the RCPSP problem. An experiment is conducted to evaluate the performance of these algorithms based on the same number of solutions calculated using ProGen generated benchmark instances. The results indicate that the MAs with regret biased sampling rule to generate initial and restart populations outperforms the other algorithms in terms of solution quality. Keywords-RCPSP; cash availability; memetic algorithms; variable neighborhood search.