Non-preemptive preferences in multi-agent task processing

Preferences are assigned on consumable resources during task execution. We have presented here a scheme for handling such preferences in nonpreemptive allocations for two types of task in a multiagent environment with the object of finding a solution that preserves the highest number of preferences. The task types are: distributed query processing (such as a travel plan) and object creation (such as cooperative design). In both cases, convergence can be guaranteed because of nonpreemption, but the quickest path for the best solution is harder to identify.