Arrow Research search

Author name cluster

Eithan Ephrati

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

9 papers
2 author rows

Possible papers

9

AAAI Conference 1996 Conference Paper

A Cost-Directed Planner: Preliminary Report

  • Eithan Ephrati

We present a cost-directed heuristic planning algorithm, which uses an A* strategy for node selection. The heuristic evaluation function is computed by a deep lookahead that calculates the cost of complete plans for a set of pre-defined top-level subgoals, under the (generally false) assumption that they do not interact. This approach leads to finding low-cost plans, and in many circumstances it also leads to a significant decrease in total planning time. This is due in part to the fact that generating plans for subgoals individually is often much less costly than generating a complete plan taking interactions into account, and in part to the fact that the heuristic can effectively focus the search. We provide both analytic and experimental results.

PDF Details

IJCAI Conference 1995 Conference Paper

Deriving Multi-Agent Coordination through Filtering Strategies

  • Eithan Ephrati
  • Martha E. Pollack
  • Sigalit Ur

We examine an approach to multi-agent coordination that builds on earlier work on enabling single agents to control their reasoning in dynamic environments. Specifically, we study a generalization of the filtering strategy. Where single-agent filtering means tending to bypass options that are incompatible with an agent's own goals, multi-agent filtering means tending to bypass options that are incompatible with other agents' known or presumed goals. We examine several versions of multi-agent filtering, which range from purely implicit to minimally explicit, and discuss the trade-offs among these. We also describe a series of experiments that demonstrate initial results about the feasibility of using multi-agent filtering to achieve coordination without explicit negotiation.

PDF

AAAI Conference 1994 Conference Paper

Divide and Conquer in Multi-Agent Planning

  • Eithan Ephrati

In this paper, we suggest an approach to multiagent planning that contains heuristic elements. Our method makes use of subgoals, and derived sub-plans, to construct a global plan. Agents solve their individual sub-plans, which are then merged into a global plan. The suggested approach may reduce overall planning time and derives a plan that approximates the optimal global plan that would have been derived by a central planner, given those original subgoals. We consider two different scenarios. The first involves a group of agents with a common goal. The second considers how agents can interleave planning and execution when planning towards a common, though dynamic, goal.

PDF Details

ICAPS Conference 1994 Conference Paper

Plan Execution Motivation in Multi-agent Systems

  • Eithan Ephrati
  • Motty Perry
  • Jeffrey S. Rosenschein

In this paper we analyse a particular modelof control amongintelligent agents, that of non-abao|ute n control. Non-absolutecontrol involves a %upervisor nagent that issues orders to a group of "subordinate agents. An example might be aa Internet user who issues a query to a group of software agents on remote hosts, or ffi humanagent on Earth directing the activities of Mars-basedsemi-autonomousvehicles. The membersof the subordinate group are assumed to be self-motlv~ted, and individually rational (i. e., they are basically willing to carry out the supervisor’s request if properly compensated). This assumption gives rise to the need for a reward policy that wouldmotivate each agent to contribute to the groupactivity. In this paper we introduce such a policy under certain simplifying assumptions.

Details

AAAI Conference 1992 Conference Paper

Constrained Intelligent Action: Planning Under the Influence of a Master Agent

  • Eithan Ephrati

In this paper we analyze a particular model of control among intelligent agents, that of non-absolute control. Non-absolute control involves a “supervisor” agent that issues orders to a “subordinate” agent. An example might be a human agent on Earth directing the activities of a Mars-based semi-autonomous vehicle. Both agents operate with essentially the same goals. The subordinate agent, however, is assumed to have access to some information that the supervisor does not have. The agent is thus expected to exercise its judgment in following orders (i. e. , following the true intent of the supervisor, to the best of its ability). After presenting our model, we discuss the planning problem: how would a subordinate agent choose among alternative plans? Our solutions focus on evaluating the distance between candidate plans.

PDF Details