Social Network-Oriented Learning Agent for Improving Group Intelligence Coordination

Abstract

Group coordination is embedded in social networks, which aims to reach a consensus or solve conflicts among social individuals. Recently, improving performance has become a challenging task and drawn considerable interest in this field. To eliminate barriers of group coordination, we abstract the problem into a networked color coordination game and introduce learning agents encoded with Q-learning to collect local information and learn local individual behaviors. We first show that learning agents can effectively improve the group coordination performance. By properly selecting parameters, we find that learning agents acting with low greedy parameter levels and placed in central locations can drastically accelerate group coordination. Greedy parameters and positions are vital to the problem. Finally, we indicate that learning agents have a direct effect on neighbor individuals and an indirect effect on non-neighbor individuals to act on the coordination network. Moreover, we propose a conflict relationship index which is the average rounds required for solving conflicts and indicate learning agents solve conflicts that cannot be solved by an individual. Hence, learning agents create further benefits to group coordination in these complex social networks. This paper provides a detailed analysis of the learning agents in a networked color coordination game and shows that artificial intelligence provides a solution to the group coordination problem. INDEX TERMS Group coordination, learning agents, reinforcement learning.

Random choices facilitate solutions to collective network coloring problems by artificial agents

“…We incorporate random behavior in various decision stages in the following modified update rules based on the basic greedy update rule above. Notably, these simple yet natural update rules based on intuitive heuristics, combined with the bipartite network structure which simplifies the possible colorings, enable analytical insights that are unobtainable in the more complicated systems put forward in other work ( Qi et al., 2019 ).…”

“…However, they also found that the bots could be detrimental if not properly tuned with the appropriate levels of randomness ( Shirado and Christakis 2017 ). Along this line, a recent related modeling work has incorporated reinforcement learning algorithms (q-bots) into agent-based simulations of the distributed coloring problem ( Qi et al., 2019 ). Despite these developments, there still is a lack of analytical insights into the optimal level of random behavior needed when solving network coloring problems.…”

How committed individuals shape social dynamics: A survey on coordination games and social dilemma games

“…They also found that the impact of bots with small noise was comparable to assigning nodes with fixed colors, directly corresponding to committed individuals, to achieve compatibility with a global solution. Other related numerical or theoretical solutions to this problem include reinforcement learning algorithms (q-bots) or myopic artificial agents [33,34]. Although these studies employed more sophisticated algorithms than simply designing agents with colors, the notion of committed individuals persists, as these preprogrammed bots maintained consistent strategies without adaptation.…”

Group Chase and Escape of Biological Groups Based on a Visual Perception-Decision-Propulsion Model

“…Related topics include shoals of fish [2]- [8], flocks of birds [9], [10], swarms of locusts [11]- [13], communities of bacteria [14], [15], groups of microtubules [16], masses of histiocytes [17], [18], streams of traffic and flows of humans [19], [20]. A relatively simple and local interaction between individuals produces coordinated and ordered collective behaviors such as these [21]. In this way, biological groups show various intelligent characteristics (distribution, self-adaption, and robustness) that cannot be achieved by a single individual in all kinds of situations.…”