This paper presents an approach for the multiagent navigation and conflict resolution problem, that considers the issue of performance. We present a decentralized predictive navigation scheme that combines the Decentralized Navigation Functions methodology with the Model Predictive Control (MPC) framework while preserving the former's collision avoidance and convergence guarantees. Aircrafts, flying at constant altitude, are modeled as unicycles. Performance criteria are encoded in a cost functional. Due to decentralization, each agent does not take into account the decisions of others in the control law calculation, resulting in performance discrepancies. Therefore, we employ event-triggered executions in our scheme. The improved performance is demonstrated through simulations.