Rescue teams require exchanging tactical information to efficiently adapt their behavior to the particular disaster response scenario conditions. Broadcasting is the main technique used to disseminate messages throughout nodes in ad hoc networks. Ad hoc networks are one of the most appealing networks to be used in disaster response scenario due to their decentralized and self-organizing features. While many broadcasting techniques have been presented for ad hoc networks, there are not adaptive broadcasting schemes for disaster response scenarios so far. The complexity of designing broadcasting protocols for disaster response scenarios is high due to the large number of factors affecting the communications among nodes, such as mobility, density, message collisions, etc. In this paper we propose an evolutionary computational approach for designing and optimizing a probability-based broadcast scheme using topological features of nodes. The validity of the proposed approach is shown through simulations.