Decision-making in non-cooperative games with conflicting self-objectives

This paper concerns multicriteria decision making by players in a conflict situation. The considered situation is modelled as a non-cooperative game. Moreover, the player has a conflict not only with the opponent, but also she is faced with her own conflicting objectives. Hence, selecting a strategy is dependent on both objective preferences and the inherent uncertainty about the opponent selection. In contrast to most studies on such multiobjective games (MOGs), which employed a scalarization approach, it has recently been suggested to pose such games as MOGs with undecided objective preferences. As in Pareto optimization, the novel solution method to the considered MOGs reveals tradeoff information, which is commonly hindered when using a classical scalarization approach. To complement and highlight the significance of the aforementioned approach, two decision analysis techniques are suggested. The proposed techniques are compared with weighted-sum utility-based analysis, as well as with the Pareto-optimal security strategy approach. The comparison is done using a MOG with two competing travelling salesperson. It is concluded that the suggested techniques allow incorporating tradeoff information to justify the selection of a strategy.