Optimization-based decision support for scenario analysis in electronic sourcing markets with volume discounts

•We consider supplier selection problems in e-sourcing with volume discounts.•We propose an optimization model which allows for effective scenario analysis.•The model reduces the number of parameter settings to explore considerably.•We provide experimental results using exact algorithms and heuristics.•We find that meta-heuristics can be effectively used for these problems.

E-Sourcing software has become an integral part of electronic commerce. Beyond the use of single-lot auction formats, there has been an emerging interest in using e-sourcing software for complex negotiations. Procurement markets typically exhibit scale economies leading to various types of volume discounts which are in wide-spread use in practice. The analysis of bids in such negotiations typically leads to computationally hard optimization problems. Scenario analysis describes a process, in which procurement managers compute different award allocations as a result of different allocation constraints and parameters that they put in place. This paper discusses an optimization model and computational methods which allow for effective scenario analysis with allocation problems in the presence of different types of discount policies and allocation constraints. The model reduces the number of parameter settings to explore considerably. The models are such that they can often not be solved exactly for realistic problem sizes in practically acceptable time frames. Therefore, we provide results of numerical experiments using exact algorithms and heuristics to solve the problem. We find that RINS and Variable Neighborhood Search can be effectively used in traditional branch-and-cut algorithms for this problem. Overall, new computational approaches allow procurement managers to evaluate offers even in markets with a complex set of volume discounts and multiple allocation constraints.