Abstract
An agent that adopts a commitment to another agent should act so as to bring about a state of the world meeting the specifications of the commitment. Thus, by faithfully pursuing a commitment, an agent can be trusted to make sequential decisions that it believes can cause an intended state to arise. In general, though, an agent’s actions will have uncertain outcomes, and thus reaching an intended state cannot be guaranteed. For such sequential decision settings with uncertainty, therefore, commitments can only be probabilistic. We propose a semantics for the trustworthy fulfillment of a probabilistic commitment that hinges on whether the agent followed a policy that would be expected to achieve an intended state with sufficient likelihood, rather than on whether the intended state was actually reached. We have developed and evaluated algorithms that provably operationalize this semantics, with different tradeoffs between responsiveness and computational overhead. We also discuss opportunities and challenges in extending our proposed semantics to richer forms of uncertainty, and to other agent architectures besides the decision-theoretic agents that have been our initial focus of study. Finally, we consider the implications of our semantics on how trust might be established and confirmed in open agent systems.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
- 2.
Because the model is assumed accurate, the agent can be assumed to only formulate policies (and thus commitments) that it is capable of executing. Permitting inaccurate models (where an agent might make a commitment it is inherently incapable of fulfilling) is outside the scope of the focus of this paper on trustworthy fulfillment of commitments.
- 3.
Recall that a policy is defined over all (belief) states, and so covers every possible contingency that could arise during execution. We refer to a particular sequence of states and (policy-dictated) actions that might be experienced as a trajectory. Note that a policy thus differs from a plan, which is typically defined in terms of a specific (nominal) trajectory. Hence, a plan can fail (stimulating plan repair or replanning) when unintended action outcomes or external events cause a deviation from the plan’s nominal trajectory. In contrast, a policy never “fails” because it specifies actions for every state (and thus for every possible trajectory).
References
Agotnes, T., Goranko, V., Jamroga, W.: Strategic commitment and release in logics for multi-agent systems (extended abstract). Technical report IfI-08-01, Clausthal University (2007)
Al-Saqqar, F., Bentahar, J., Sultan, K., El-Menshawy, M.: On the interaction between knowledge and social commitments in multi-agent systems. Appl. Intell. 41(1), 235–259 (2014)
Bannazadeh, H., Leon-Garcia, A.: A distributed probabilistic commitment control algorithm for service-oriented systems. IEEE Trans. Netw. Serv. Manage. 7(4), 204–217 (2010)
Castelfranchi, C.: Commitments: from individual intentions to groups and organizations. In: Proceedings of the International Conference on Multiagent Systems, pp. 41–48 (1995)
Chesani, F., Mello, P., Montali, M., Torroni, P.: Representing and monitoring social commitments using the event calculus. Auton. Agent. Multi-Agent Syst. 27(1), 85–130 (2013)
Cohen, P.R., Levesque, H.J.: Intention is choice with commitment. Artif. Intell. 42(2–3), 213–261 (1990)
Durfee, E.H., Singh, S.: Commitment semantics for sequential decision making under reward uncertainty. In: Papers from the AAAI Fall Symposium on Sequential Decision Making for Intelligent Agents (AAAI Technical report FS-15-06) (2015)
Jennings, N.R.: Commitments and conventions: the foundation of coordination in multi-agent systems. Knowl. Eng. Rev. 8(3), 223–250 (1993)
Kinny, D.N., Georgeff, M.P.: Commitment and effectiveness of situated agents. In: Proceedings of the 1991 International Joint Conference on Artificial Intelligence (IJCAI-91), pp. 84–88 (1991)
Maheswaran, R., Szekely, P., Becker, M., Fitzpatrick, S., Gati, G., Jin, J., Neches, R., Noori, N., Rogers, C., Sanchez, R., Smyth, K., Buskirk, C.V.: Look where you can see: predictability & criticality metrics for coordination in complex environments. In: International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS) (2008)
Mallya, A.U., Huhns, M.N.: Commitments among agents. IEEE Internet Comput. 7(4), 90–93 (2003)
Poupart, P., Vlassis, N., Hoey, J., Regan, K.: An analytic solution to discrete Bayesian reinforcement learning. In: Proceedings of International Conference on Machine Learning (ICML 2006), pp. 697–704 (2006)
Raffia, H.: The Art and Science of Negotiation. Harvard University Press, 79 Garden St. (Belknap Press) (1982)
Ramachandran, D., Amir, E.: Bayesian inverse reinforcement learning. In: International Joint Conference on Artificial Intelligence (IJCAI), pp. 2586–2591 (2007)
Sandholm, T., Lesser, V.R.: Leveled commitment contracts and strategic breach. Games Econ. Behav. 35, 212–270 (2001)
Singh, M.P.: An ontology for commitments in multiagent systems. Artif. Intell. Law 7(1), 97–113 (1999)
Singh, M.P.: Commitments in multiagent systems: some history, some confusions, some controversies, some prospects. The Goals of Cognition. Essays in Hon. of C. Castelfranchi, pp. 1–29 (2012)
Vokrínek, J., Komenda, A., Pechoucek, M.: Decommitting in multi-agent execution in non-deterministic environment: experimental approach. In: 8th International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS 2009), pp. 977–984 (2009)
Winikoff, M.: Implementing flexible and robust agent interactions using distributed commitment machines. Multiagent Grid Syst. 2(4), 365–381 (2006)
Witwicki, S., Chen, I.T., Durfee, E., Singh, S.: Planning and evaluating multiagent influences under reward uncertainty (extended abstract). In: Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS), pp. 1277–1278 (2012)
Witwicki, S., Durfee, E.: Commitment-based service coordination. Int. Jour. of Agent-Oriented Softw. Eng. 3(1), 59–87 (2009)
Xing, J., Singh, M.P.: Formalization of commitment-based agent interaction. In: Proceedings of the 2001 ACM Symposium on Applied Computing (SAC), pp. 115–120 (2001)
Xuan, P., Lesser, V.R.: Incorporating uncertainty in agent commitments. In: Jennings, N.R., Lespérance, Y. (eds.) ATAL 1999. LNCS (LNAI), vol. 1757, pp. 57–70. Springer, Heidelberg (2000). doi:10.1007/10719619_5
Acknowledgments
Inn-Tung (Anna) Chen, Stefan Witwicki, Alexander Gutierrez, and Qi Zhang contributed to various aspects of the ideas and algorithms described in this paper. We also benefited from detailed discussions about commitments with Munindar Singh and Scott Gerard, and from the questions and suggestions of the anonymous reviewers. This work was supported in part by the Air Force Office of Scientific Research under grant FA9550-15-1-0039.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing AG
About this paper
Cite this paper
Durfee, E.H., Singh, S. (2016). On the Trustworthy Fulfillment of Commitments. In: Osman, N., Sierra, C. (eds) Autonomous Agents and Multiagent Systems. AAMAS 2016. Lecture Notes in Computer Science(), vol 10002. Springer, Cham. https://doi.org/10.1007/978-3-319-46882-2_1
Download citation
DOI: https://doi.org/10.1007/978-3-319-46882-2_1
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-46881-5
Online ISBN: 978-3-319-46882-2
eBook Packages: Computer ScienceComputer Science (R0)