ChargingBoul: A Competitive Negotiating Agent with Novel Opponent Modeling

ChargingBoul: A Competitive Negotiating Agent with Novel Opponent Modeling Joe Shymanski December 9, 2025 Abstract Automated negotiation has emerged as a critical area of research in multiagent systems, with applications spanning e-commerce, resource allo- cation, and autonomous decision-making. This paper presents Charging- Boul, a negotiating agent that competed in the 2022 Automated Negoti- ating Agents Competition (ANAC) and placed second in individual utility by an exceptionally narrow margin. ChargingBoul employs a lightweight yet effective strategy that balances concession and opponent modeling to achieve high negotiation outcomes. The agent classifies opponents based on bid patterns, dynamically adjusts its bidding strategy, and applies a concession policy in later negotiation stages to maximize utility while fos- tering agreements. We evaluate ChargingBoul’s performance using com- petition results and subsequent studies that have utilized the agent in ne- gotiation research. Our analysis highlights ChargingBoul’s effectiveness across diverse opponent strategies and its contributions to advancing au- tomated negotiation techniques. We also discuss potential enhancements, including more sophisticated opponent modeling and adaptive bidding heuristics, to improve its performance further. 1 Introduction Automated negotiation has long been a critical area of research in artificial intel- ligence, with applications ranging from e-commerce and diplomatic mediation to resource allocation in multiagent systems. The Automated Negotiating Agents Competition (ANAC) serves as a benchmark for advancing the field, challenging participants to develop autonomous agents capable of strategic, adaptive, and efficient negotiation. In 2022, we entered ChargingBoul into the Automated Negotiation League (ANL) of ANAC, where it competed against dozens of agents submitted by teams from around the world. The competition featured bilateral multiattribute negotiations, where each agent sought to maximize its utility while interacting with diverse opponents. Each negotiation session consisted of 50 sequential, timed rounds, with agents alternately exchanging bids based on predefined pref- 1 arXiv:2512.06595v1 [cs.MA] 6 Dec 2025 erence profiles. The final agreement, if reached before time expired, determined the agents’ respective utilities, calculated via a linear additive utility function. Performance in the competition was assessed using two primary metrics: 1. Individual Utility: The average utility an agent secured across negoti- ations. 2. Social Welfare: The sum of utilities between both negotiating agents per round. ChargingBoul proved to be highly competitive, securing second place in in- dividual utility, falling short of first place by only 0.001 utility points on average. Its success stemmed from a straightforward yet effective negotiation strategy, leveraging opponent modeling and adaptive bidding techniques. This paper de- tails the design choices that contributed to ChargingBoul’s performance, com- pares its results to previous approaches, and discusses potential improvements for future autonomous negotiating agents. 2 Literature Review Automated negotiation has been widely studied within the domains of multi- agent systems and game theory. Early approaches to autonomous negotiation were inspired by classical economic models, such as Nash bargaining solutions and concession-based tactics. More recent research has emphasized adaptive strategies that leverage opponent modeling, machine learning, and utility-driven decision-making. The ANAC competition has provided a valuable testing ground for these advancements, with previous studies analyzing both individual agent strategies and broader trends in competitive negotiation. Notable approaches include de Jonge’s MiCRO strategy [3], the ANOTO agent [2], and Luckyagent2022 [4], which have demonstrated the effectiveness of heuristic-based bidding, learning- based adaptation, and opponent classification techniques. Like many of its predecessors, ChargingBoul builds upon existing method- ologies, employing a structured opponent modeling framework alongside a dy- namically adaptive strategy. Additionally, ChargingBoul introduces two novel statistics by which it classifies opponent negotiation tactics and subsequently adjusts its bidding strategy. By evaluating its performance within ANAC 2022, this paper contributes to the ongoing discourse on effective autonomous negotia- tion and explores how lightweight yet effective strategies can achieve competitive results. 3 Methodology An automated negotiation agent typically contains three main components: its own bidding strategy, an acceptance strategy, and a method for modeling the op- ponent’s bidding strategy. ChargingBoul implements all three. We will present 2 Figure 1: Overview of ChargingBoul’s entire negotiation strategy. its novel opponent modeling first to introduce a few key terms that are neces- sary to understand th

This content is AI-processed based on open access ArXiv data.