580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How can negotiation protocols and frameworks be designed to structure and govern automated e-commerce negotiations effectively?
This research area focuses on formalizing negotiation processes in e-commerce by developing and analyzing negotiation protocols and frameworks that define negotiation states, permissible actions, and communication rules. Proper protocol design is crucial to ensure flexible yet structured negotiation, coordination of multi-agent interactions, compatibility with human negotiators, and adaptability to diverse negotiation types (e.g., auctions, bilateral, multilateral). Such protocols underpin the automation of negotiation, supporting both agents' internal reasoning and inter-agent communication for efficient and reliable e-commerce transaction settlements.
Key finding: This paper models various business negotiation mechanisms, including fixed price sales, auctions, and brokerages, within a unified finite state machine (FSM) framework, revealing structural commonalities in e-commerce... Read more
Key finding: The paper elucidates the theoretical foundations for designing negotiation protocols that balance flexibility for human users and automated system enforcement. It proposes a negotiation protocol formalism that tightly couples... Read more
Key finding: This work addresses limitations of existing online auction mechanisms where bidders learn payments only upon departure, defining prompt mechanisms that disclose payments immediately upon winning. It designs deterministic and... Read more
Key finding: The proposed Negotiation Situation Specification Scheme (N3S) framework structures both internal cognitive negotiation processes and external communication via argumentation. By identifying basic argument types and relating... Read more
Key finding: This paper extends classical alternating-offers bargaining models by allowing negotiation without full preference disclosure, strategic delay signaling, proposal revisions, and belief-changing arguments. Introducing eight... Read more
2. What negotiation strategies and agent decision-making models optimize outcomes in automated e-commerce negotiations under uncertainty and multiple issues?
This thematic area explores agent-based negotiation strategies that guide proposal generation, concession tactics, opponent modeling, and multi-issue tradeoff handling to optimize negotiation utility. Emphasis is on dynamic, adaptive strategies informed by utilities, temporal constraints, behavioral learning, and multi-attribute preferences. Effective negotiation strategies enable agents to achieve higher agreement quality and faster convergence in complex e-commerce settings involving multiple parties and issues, critical for practical deployment.
Key finding: The study presents a taxonomy of negotiation strategies incorporating concession tactics and alternative search tactics tailored for mobile agent negotiation in e-trade. It emphasizes adaptive strategy selection based on... Read more
Key finding: This paper introduces a two-level negotiation process where the upper level sets high-level goals and decides engagement, while the lower level refines commitments considering feasibility and constraints. This hierarchical... Read more
Key finding: The approach integrates motivational qualities (MQ) framework for multi-goal evaluation, partial-order scheduling for concurrent multi-linked negotiation issues, and multi-level negotiation abstraction. Agents reason about... Read more
Key finding: The paper proposes a heuristic model for agents participating in multiple concurrent bilateral negotiations under incomplete information and deadlines. It coordinates negotiation threads mutually influencing bidding stances... Read more
Key finding: Besides protocol modeling, the paper also elaborates on the finite-state machine representation of negotiation messages and states that underpin negotiation strategies for various auction types and brokered sales, helping... Read more
3. How do automated negotiation systems compare with human negotiation, and what decision support or computational methods enhance negotiation outcomes in e-commerce?
This theme covers empirical and simulation-based comparative studies of automated negotiation systems versus human negotiators, investigating transaction cost reductions, outcome quality, and efficiency benefits of automation. It also encompasses decision support mechanisms, utility-based evaluation, and post-negotiation optimization procedures that compensate for human biases and power asymmetries. Understanding these dynamics aids in designing automated negotiation frameworks that can outperform or augment human negotiation in e-commerce.
Key finding: By modeling and simulating automated negotiation systems with concession strategies and comparing to outcomes from over 2000 human negotiation experiments, the study reveals that automated systems reduce transaction costs... Read more
Key finding: The study demonstrates that mathematical post-negotiation procedures, specifically the Adjusted Winner (AW) algorithm, can significantly improve negotiation efficiency and fairness, even in power-asymmetric situations common... Read more
Key finding: This research develops a Bayesian network-based technique to elicit and model users' valuations of private information disclosure consequences. By reducing uncertainty through utility elicitation and defining optimal stopping... Read more
Key finding: The PrivacyPact protocol models multi-issue negotiation exchanges of private information for rewards, ensuring completeness and convergence when mutual agreements exist. Utility theory guides participants in making offers and... Read more
Key finding: The paper introduces a law-governed interaction mechanism that separates policy definition ('law') from enforcement via trusted controllers. This formalism enables unified, secure enforcement of diverse e-commerce policies... Read more