1 / 23

Trust Based Mechanism Design

Trust Based Mechanism Design. Use MD. Motivation. Fuse the fields of trust-modelling and mechanism design Trust measures how good an interaction partner is Mechanism design concerns itself with the allocation of resources. . i. i. p. p. r. r. c. c. e. e. Y. X.

sammy
Download Presentation

Trust Based Mechanism Design

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Trust Based Mechanism Design

  2. Use MD

  3. Motivation • Fuse the fields of trust-modelling and mechanism design • Trust measures how good an interaction partner is • Mechanism design concerns itself with the allocation of resources.

  4. i i p p r r c c e e Y X Deciding on a service provider Garage X is better than Y X Garage Y is better than X! Y

  5. Motivation WHICH GARAGE TO EMPLOY? X Y

  6. Mechanism Design • The design of systems of interacting agents • Protocols (allocation + payment schemes) to ensure that certain global properties are achieved at equilibrium (game-theoretic). • Selfish, rational agents possessing private information i.e. each agent maximises its utility function given its type.

  7. Task Allocation • Agent 4 wants a task to be completed. • Problem • Who to allocate task to ? • How much to pay task performer ? • Solution • VCG mechanism

  8. Applying VCG Centre (£210,task1) Allocate task to 1, 4 pays 1 £50 4 Ask(£40,task1) Ask(£50,task1) Ask(£80,task1) 1 2 3

  9. VCG Mechanism • Class of mechanisms satisfying • Efficiency – outcome maximising utility for all agents • Individually rationality – incentivise participation • Incentive compatibility – incentivise truthful report under dominant strategy: • Works by: • Allocation : choose allocation maximising utility • Payment : charge marginal utility contributed by agent

  10. Cheapest not always best …..

  11. i [ ( ) ] µ E C K ´ v i 4 4 4 ; Adding uncertainties • Suppose that agent 4 has formulated a belief about the Probability of Success (POS) of other agents.

  12. Trust • So far : one’s own experience. • Would be better to factor in others’ experiences as well. • First time agent • Greater pool of experience for seasoned user. • Trust a way of aggregating everyone’s experience

  13. Our model of trust • Each time a task is performed, record POS • Publicly transmit the POS to other agents • Trust • Use a simple averaging function over reports from all agents • Give more weight to more trusted referrals, or agents with similar properties 4 2 3 1 Record POS Transmit POS

  14. [ ( [ ) ] ] ( ) µ E j 0 3 0 2 0 1 0 4 2 1 0 1 ® v v ¿ = = 4 4 4 ; : : : : ; i 3 2 1 t ´ ´ ´ 4 i i i c i Incorporate Trust VCG Fails!!

  15. b ¤ K 5 8 4 0 0 0 c c c = = = 1 3 2 i i i i ´ ´ ´ ´ 4 2 3 1 Centre calculates optimal allocation TBMD (£210,task1) Trust model Centre 4 3 1 2 4 transmits to centre -- value of task, observed POS and trust model 1, 2, 3 transmit to centre -- cost of doing task and observed POS

  16. b ( ) ( ) ¤ ¤ ¤ K D U K U K ¡ = i i i ¡ ¡ 8 4 5 0 0 0 ; : ; : c c c = = = 3 2 1 i i i i ´ ´ ´ ´ 4 1 2 3 Centre calculates optimal allocation without each agent i’s POS report. TBMD (£210,task1) Trust model Centre 4 3 1 2

  17. D ¡ r m c = i i i 4 8 5 0 0 0 c c c = = = 2 1 3 ¤ K i i i i r i ; ´ ´ ´ ´ 3 2 4 1 Centre calculates transfer to i , TBMD (£210,task1) Trust model Centre 4 3 1 2

  18. Steps in TBMD

  19. Trust Based Mechanism Design (TBMD) • An individually rational and efficient ICDR mechanism that can also select those agents that are most successful at their task

  20. Experimental Setup • Set of buyers and sellers. • Biased (but truthful!) report from seller. • Buyers attach different levels of importance to seller’s report.

  21. Results K* K*TBM K*TBM0.25 K*VCG K*TBM0.5 K*FTM

  22. Conclusions • Task allocation with uncertainty. • Only incentive-compatible, individually-rational efficient reputation mechanism. • Works with a very large class of trust models.

  23. End rkd02r@ecs.soton.ac.uk Any Questions?

More Related