The Radar Resource Management (RRM) problem for a multi-target tracking application is considered. The problem is defined as a stochastic optimal control problem with resource constraints. A novel Bayes risk-based formulation for the objective function, with the aim of allocating the radar resources by minimizing the risk of committing track maintenance errors, is proposed. The formulation proposed increases the interpretability in the decision-making process and conforms more with the system user's wishes. The effectiveness and benefits of active time resources management in dynamic conditions are demonstrated by means of simulations.