Aggregating Correlated Wind Power with Full Surplus Extraction

We study the problem of designing profitmaximizing mechanisms for an aggregator who aggregates wind power from a group of wind power producers (WPPs). The WPPs have more refined forecasts of the wind power generation than the aggregator. Such forecasts are their private information, which also give the reservation utilities of the WPPs. The goal of the aggregator is to elicit the private information truthfully, while paying them as little as possible. Inspiring by the fact that those forecasts are typically correlated due to the geographical proximity of the WPPs, we formally define the full correlation condition, which holds ubiquitously in practice. Under that condition, we construct an optimal mechanism which yields the truthful elicitation, while extracting the full surplus (i.e., with minimum payments equal to the reservation utilities) in expectation. Finally, we conduct a case study based on the realworld data, which empirically validates the results.