Scenario Tree Generation Using Data Reconciliation For Aggregators Bidding Strategy

For multistage decision-making for an uncertain future, developing a model that captures the interrelation between the underlying uncertain parameters of different stages is crucial. To this end, scenario tree approaches are widely applied. The success of a scenario tree approach relies on 1) utilising as much data as available and 2) integrating the effect of current stage parameters in the next stage(s) parameters. A novel scenario tree generation approach based on data reconciliation is developed in this study to achieve both. A linear programming moment-matching method is then adopted to alleviate the computational burden, making it viable to represent the state space with significantly fewer scenarios. The proposed method is applied in determining the bidding strategy of aggregators in energy markets to showcase its effectiveness in real-life decision-making applications. Distribution System Aggregators (DSAs) should incorporate market price uncertainty and multiple market interrelations. For this purpose, the multistage scenario tree method is effective. The results of applying the proposed approach are compared with those when data reconciliation is not applied and when a scenario tree is not employed. The comparison explicitly indicates the effectiveness of the proposed scenario tree generation technique, as it yields the greatest return for the DSA.