In multi-objective multi-agent systems (MOMASs), agents explicitly consider the possible trade-offs between conflicting objective functions. We argue that compromises between competing objectives in MOMAS should be analyzed on the basis of the utility that these compromises have for the users of a system, where an agent’s utility function maps their payoff vectors to scalar utility values. This utility-based approach naturally leads to two different optimization criteria for agents in a MOMAS: expected scalarized returns (ESRs) and scalarized expected returns (SERs). In this article, we explore the differences between these two criteria using the framework of multi-objective normal-form games (MONFGs). We demonstrate that the choice of optimization criterion (ESR or SER) can radically alter the set of equilibria in a MONFG when nonlinear utility functions are used.

Many real-world domains can benefit from adaptable decentralized task allocation through emergent specialization, especially in large teams of non-communicating agents. We begin with an existing bio-inspired response threshold reinforcement approach for decentralized task allocation and extend it to handle hierarchical task domains. We test the extension on self-deployment of a large team of non-communicating agents to patrolling a hierarchically defined set of areas. Results show near-ideal performance across all areas, while minimizing wasteful task switching through the development of specializations and subsequent respecializations when area demands change. A genetic algorithm is then used to evolve even more adaptable and stable task allocation behavior, by incorporating weight and power coefficients into agents’ response threshold reinforcement action probability calculations.

Recent progress in reinforcement learning (RL) using self-play has shown remarkable performance with several board games (e.g., Chess and Go) and video games (e.g., Atari games and Dota2). It is plausible to hypothesize that RL, starting from zero knowledge, might be able to gradually approach a winning strategy after a certain amount of training. In this paper, we explore neural Monte Carlo Tree Search (neural MCTS), an RL algorithm that has been applied successfully by DeepMind to play Go and Chess at a superhuman level. We try to leverage the computational power of neural MCTS to solve a class of combinatorial optimization problems. Following the idea of Hintikka’s Game-Theoretical Semantics, we propose the Zermelo Gamification to transform specific combinatorial optimization problems into Zermelo games whose winning strategies correspond to the solutions of the original optimization problems. A specially designed neural MCTS algorithm is then introduced to train Zermelo game agents. We use a prototype problem for which the ground-truth policy is efficiently computable to demonstrate that neural MCTS is promising.

There is an implicit assumption in machine learning techniques that each new task has no relation to the tasks previously learned. Therefore, tasks are often addressed independently. However, in some domains, particularly reinforcement learning (RL), this assumption is often incorrect because tasks in the same or similar domain tend to be related. In other words, even though tasks are quite different in their specifics, they may have general similarities, such as shared skills, making them related. In this paper, a novel domain adaptation-based method using adversarial networks is proposed to do transfer learning in RL problems. Our proposed method incorporates skills previously learned from source task to speed up learning on a new target task by providing generalization not only within a task but also across different, but related tasks. The experimental results indicate the effectiveness of our method in dealing with RL problems.