Efficient algorithms for robust procurements

Starting with a project and several agents willing to execute it, the goal is to find a sequence of agents which can work on the project if the previous agent fails. Each sequence has a total cost and probability of finishing within the deadline. Two objectives for optimization are introduced: reliability and cost. We model a bi-objective optimization problem for solving these procurements efficiently. First, we propose an alternative for the traditional weighted sum method to find a Pareto front. Furthermore, an adapted branch and bound algorithm is proposed to compete with the adaptive weighted sum method in terms of quality and computation time. In the experiments we present the running times on which the number of agents has huge influence. The discussion handles the shape of the Pareto front and discusses its dependence on the parameters. In terms of both cost and time efficiency, our algorithm offers a useful tool for solving procurement problems