This paper considers the problem of distributed dynamic task allocation by a set of cooperative agents. There are different types of tasks that are dynamically arriving to a system. Each of the agents can satisfy the tasks with some quality (which may be zero). Every task is augmented by the needed qualitative level of task’s execution. Thus, relation between agents and task types is fuzzy. The main goal of the agents is to maximize the overall performance of the system and to fulfill the tasks as soon as possible. This problem belongs to the Distributed Problem Solving class of Distributed Artificial Intelligence research. The results differ from that for task allocation in multi-agent environments where each agent tries to maximize its own performance.