Sensitivity Analysis of Uncertainties in Dynamic Noncooperative Games Using a Multi-Modeling Method

In this paper, we formulate a two-person nonzero-sum game with unknown parameters by multiple models and carry out a sensitivity analysis of performance indices with respect to the unknowns. We show that from each player’s point of view, the deviation from the indices is bounded from below and above if he/she adopts his/her estimates of the uncertain parameters while assuming the competitors use unknown values of the parameters. Motivated by the model for strategic bidding in a competitive electricity market, performance indices with linear and cross terms in control have been constructed and a simple numerical example is given for illustration purpose.