Improving Performance of Heterogeneous Agents

With the increase in agent-based applications, there are now agent systems that support concurrent client accesses. The ability to process large volumes of simultaneous requests is critical in many such applications. In such a setting, the traditional approach of serving these requests one at a time via queues (e.g. FIFO queues, priority queues) is insuucient. Alternative models are essential to improve the performance of such heavily loaded agents. In this paper, we propose a set of cost-based algorithms to optimize and merge multiple requests submitted to an agent. In order to merge a set of requests, one rst needs to identify commonalities among such requests. First, we provide an application independent framework within which an agent developer may specify relationships (called invariants) between requests. Second, we provide two algorithms (and various accompanying heuristics) which allow an agent to automatically rewrite requests so as to avoid redundant work|these algorithms take invariants associated with the agent into account. Our algorithms are independent of any speciic agent framework. For an implementation, we implemented both these algorithms on top of a geographic database agent. Based on these implementations, we conducted experiments and show that our algorithms are considerably more eecient than methods that use the A algorithm. A heavily loaded agent is one that experiences a large volume of service requests and/or has a large number of conditions to track on behalf of various users. The traditional model for servicing requests is via one kind of queue or the other (e.g. FIFO, LIFO, priority queue, etc.). For instance, a company may deploy a PowerPoint agent ppt that automatically creates PowerPoint presentations for diierent users based on criteria they have registered earlier. The nance director may get the latest The rst and third authors gratefully acknowledge support from the Army Research Laboratory under contract number DAAL01-97-K0135, and by DARPA/AFRL under grant number F306029910552.