New market-based decentralized algorithms are proposed for the task assignment of multiple unmanned aerial vehicles in dynamic environments with a limited communication range. In particular, a cooperative timing mission that cannot be performed by a single vehicle is considered. The baseline algorithms for a connected network are extended to deal with time-varying network topology including isolated subnetworks due to a limited communication range. The mathematical convergence and scalability analyses show that the proposed algorithms have a polynomial time complexity, and numerical simulation results support the scalability of the proposed algorithm in terms of the runtime and communication burden. The performance of the proposed algorithms is demonstrated via Monte Carlo simulations for the scenario of the suppression of enemy air defenses.