Abstract
Bidding strategy is an important part of a negotiation strategy in automated multi-issue negotiations. In order to present good offers, which help maximize the agent’s utility, we need to search the outcome space and find appropriate bids. Bid search can become challenging in large outcome spaces with more than ten thousands of possible bids. The traditional search methods such as exhaustive or binary search are not efficient enough to find the right bids in a large space. This is mostly due to the high number of issues, high number of possible values for each issue, and increased time complexity of usual search methods. In this paper, we investigate the potential of using meta-heuristic methods for optimizing bid search in large outcome spaces. We apply some of the most popular meta-heuristic algorithms for bid search in bidding strategy of baseline negotiating agents and evaluate their impacts on negotiation performance in different negotiation domains. The evaluation results obtained through comprehensive experiments show how meta-heuristic algorithms can help improve bid search capability and consequently negotiation performance of the agents on different performance criteria. In addition, we show which search algorithm is most suitable for improving any particular performance criterion.
Highlights
Automated negotiations is one of the main application areas of multi-agent systems in which intelligent and autonomous software agents negotiate on behalf of humans try to maximize their interests by taking into account their owner’s preferences (Baarslag, 2016; Marsa-Maestre et al, 2014)
We briefly introduce each of the selected meta-heuristic algorithms: Local search (Hill Climbing): It is the simplest type of search algorithm which starts with one random solution and in each iteration, a neighbor with better quality in terms of objective function replaces the current solution (Hoos & Stützle, 2004). n order to implement local search in this article, we use the best improvement method
We present the evaluation results in terms of the three experiment stages described in the previous section
Summary
Automated negotiations is one of the main application areas of multi-agent systems in which intelligent and autonomous software agents negotiate on behalf of humans try to maximize their interests by taking into account their owner’s preferences (Baarslag, 2016; Marsa-Maestre et al, 2014). If the agent cannot find the desired offer in the large outcome space due to lack of time, it may send non-optimal bids to its opponent and gain little utility in the negotiation session. The time complexity of this search is equal to O n in terms of the size of the outcome space if it includes n possible bids For small domains, this is a sensible solution, because with a simple sequential search, the desired offer is obtained at an acceptable time. Considering the time required to implement the concession tactic and calculate the target utility in each round of negotiation, the time to search and find the desired offer in a large space will be much less than this. The size of the negotiation domain (D) can be calculated as ∏ m
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
