Opponent Modeling in Agents’ Negotiations Using Issues’ Frequency

in automated negotiations, on behalf of their owners, agents try to obtain the best possible utility. But agents usually don’t want to reveal their interests. Modeling opponents allows them to offer better deals and increase negotiation performance. In this paper, a frequency-based modeling approach is introduced, which takes into account the number of times the value of a negotiation domain’s issue changes and the frequency of a value it is offered. Experimental results show that an agent using this model can acquire better individual utility and social welfare. This method was used by AgreeableAgent one of the participants of ANAC2018 competition which eventually ranked first out of 22 international participants in the individual utility rankings