Cost-effective planning of timed mobile agents

The two most significant planning factors in mobile agent planning (MAP) are the number of agents used and each agent's itinerary. These two planning factors must be well-scheduled, since badly-scheduled factors can cause longer execution times because of the higher routing costs. In addition to these two factors, the time constraints that reside on the nodes of the information repository (i.e. the information servers) also have to be dealt with. Consider the nodes that present correct information only for a certain time interval. If an agent is sent to gather information and arrives earlier than a specified update time, it may retrieve useless or corrupted information. To cope with these types of information retrieval, we propose a time-constrained MAP method which finds the minimum number of agents needed and the best scheduled agent itineraries for retrieving information from a distributed computing environment. The method works under the time constraints mentioned above, allows the completion time to be lower-bounded and minimizes routing overheads. Simulation results show that the proposed method produces results that are highly applicable to the time-constrained distributed information retrieval problem domain.