130 likes | 479 Views
Market-Driven Multi-Agent Collaboration in Robot Soccer Domain Presented by Barış Kurt Outline What is a Market-Driven Strategy? How it works? Different Implementations What is a Market-Driven Strategy? Ap pl ication of the basic properties of free market economy to increase the benefit
E N D
Market-Driven Multi-Agent Collaboration in Robot Soccer Domain Presented by Barış Kurt
Outline • What is a Market-Driven Strategy? • How it works? • Different Implementations
What is a Market-Driven Strategy? • Application of the basic properties offree market economy to increase the benefit • Based on the basic assumption that maximizing individual profits will approximate global profit maximization • Used for multi-agent task allocation
How it works? • There exists an overall goal of the team that an outside entity is assumed to offer a payoff • The goal is decomposed into smaller tasks and an auction is performed for each of them • For each task, agents offer their prices to the auctioneer • The bidder with the lowest offeredprice will be given the right of execution of the task
Prices of Tasks: The Cost Functions • The prices that agents offer for tasks are the estimated costs for accomplishing that task • For example: Cattacker = M2*distBall + M2*distOppGoal
Auction Mechanism In Cerberus 2005 if(playerNumber==minDistOwner&&playerNumber==minDistOwnerForOppGoal) { robotRole=RR_PRIMARY_DEFENDER; } elseif(playerNumber==minDistOwner) { robotRole=RR_SECONDARY_ATTACKER; } else if(playerNumber==minDistOwnerForOppGoal) { robotRole=RR_SECONDARY_ATTACKER; } else if(playerNumber==minDistOwnerToOwnGoal) { robotRole=RR_PRIMARY_DEFENDER; } else { robotRole=RR_SECONDARY_ATTACKER; }
Different Implementations • Centralized • Distributed • Hybrid?
Centralized Implementation • There exists a master agent (auctioneer) that controls the auctions and assigns the roles. • The master agent receives offers from all other agents for each task and sends the auction results back. • Computationally efficient. • Prone to single point failures.
Distributed Implementation • No master agent. • Every agent broadcasts its offer for every task. • Every agent runs the same auction mechanism and parallely computes the auction results. • Robust against single point failures • Requires more computation in total.
Hybrid Implementation • There exists a master agent • There is also an auction for the task of being the master • Robust against single point failures • Computationly efficient • Still not implemented, no test results.
Thanks.. • Questions?