AbstractThis paper describes a group problem solving method called 1TM (Iterated Task Migration) for dynamic task allocation problems. ITM initially assigns unoptimal task allocations to agents. Then the agents try to change their task allocations, toward the globally optimal allocation, by iteration of task migrations. The cooperation by the agents results in adaptability and parallelism, which contribute to solving the dynamic task allocation problems. ITM works in two ways, in a synchronized way (called ITM-S) and in an asynchronized way (called ITM-A). To evaluate ITM-S and ITM-A, we have applied them to the trouble recovery problem in the area of transportation. The experimental results show that ITM-S improves adaptability while ITM-A improves run time.