Efficient bidding strategies for Cliff-Edge problems

Abstract

In this paper, we propose an efficient agent for competing in Cliff-Edge (CE) and simultaneous Cliff-Edge (SCE) situations. In CE interactions, which include common interactions such as sealed-bid auctions, dynamic pricing and the ultimatum game (UG), the probability of success decreases monotonically as the reward for success increases. This trade-off exists also in SCE interactions, which include simultaneous auctions and various multi-player ultimatum games, where the agent has to decide about more than one offer or bid simultaneously. Our agent competes repeatedly in one-shot interactions, each time against different human opponents. The agent learns the general pattern of the population's behavior, and its performance is evaluated based on all of the interactions in which it participates. We propose a generic approach which may help the agent compete against unknown opponents in different environments where CE and SCE interactions exist, where the agent has a relatively large number of alternatives and where its achievements in the first several dozen interactions are important. The underlying mechanism we propose for CE interactions is a new metaalgorithm, deviated virtual learning (DVL), which extends existing methods to efficiently cope with environments comprising a large number of alternative decisions at each decision point. Another competitive approach is the Bayesian approach, which learns the opponents' statistical distribution, given prior knowledge about the type of distribution. For the SCE, we propose the simultaneous deviated virtual reinforcement learning algorithm (SDVRL), the segmentation meta-algorithm as a method for extending different basic algorithms, and a heuristic called fixed success probabilities (FSP). Experiments comparing the performance of the proposed algorithms with algorithms taken from the literature, as well as other intuitive meta-algorithms, reveal superiority of the proposed algorithms in average payoff and stability as well as in accuracy in converging to the optimal action, both in CE and SCE problems.

Bargain strategies for agent automated negotiation in an e-business environment

“…An agent, which learns the general pattern of behavior, based on all of the interactions in which it participates is developed in [15]. A generic approach, which may help the agent compete against unknown opponents in different environments is proposed.…”

On Two-Player Interval-Valued Fuzzy Bayesian Games

“…Game theory provides a useful framework for modeling/analyze cooperation, negotiation, and coalition in multiagent systems (MAS) [2][3][4][5] and social simulation. 10 Whenever probabilities are assigned for the different types of the interacting agents, this kind of strategic interaction constitutes a Bayesian game, which is a game of imperfect information, 11 the main concern of this paper. 10 Whenever probabilities are assigned for the different types of the interacting agents, this kind of strategic interaction constitutes a Bayesian game, which is a game of imperfect information, 11 the main concern of this paper.…”

“…[6][7][8][9] There are several applications in which the interacting agents only know their own characteristics and must make decisions while having to estimate the characteristics of the other participants of the interaction, which configures a game of incomplete information, as, e.g., the case of auctions in the context of MAS, which is frequently used for resource management involving two or more agents competing for the resources, 2 or the case of Cliff-Edge problems, where maximizing profits while preventing the entire deal from falling through is a great challenge for the agents. 10 Whenever probabilities are assigned for the different types of the interacting agents, this kind of strategic interaction constitutes a Bayesian game, which is a game of imperfect information, 11 the main concern of this paper. In the case of Bayesian games, sometimes it is very difficult to characterize the private information of each agent (e.g., ability, level of effort, influence, personality, interest, strategy), to establish the probabilities of the types that each player may assume.…”