Lastly, in scenarios where multiple agents are present, the behavioral mixture of agents approach, for example Vinyals et al. (2019) samples the final agent from the Nash distribution of the set of agents, can be utilized. Given that different agents, or experts, may recommend varying actions for an identical state, this results in an intrinsic stochastic policy, taking advantage of the diversity in agent decisions. If the state space is continuous, a common approach is to transform the actions into a normal or beta distribution. We apply one-hot encoding with temperature-scaled softmax. A discrete action space can be represented as a one-hot encoded vector, For instance, if action 2 out of 5 is chosen, its one-hot representation is [0, 1, 0, 0, 0], the scale the one-hot vector to [0, 1/τ, 0, 0, 0]. The higher the temperature coefficient τ, the more spread out the distribution becomes, while a lower temperature coefficient nudges the distribution closer to a deterministic action.

Microscopic traffic simulators are powerful tools for transportation engineers and planners to analyze and predict the impact of microscopic adjustments on traffic patterns without disrupting real-world traffic. For example, it can help analyze how changing road shape like replacing an intersection with a roundabout affects traffic patterns [1], and develop traffic-aware autonomous driving policies that enhance overall traffic efficiency [2, 3]. However, creating a realistic simulator that can simultaneously replicate the microscopic response of human drivers to traffic conditions and the resulting long-term macroscopic statistics is a challenging task. In recent years, there have been significant efforts to develop realistic traffic simulators that accurately model human driving behavior. Traditional traffic simulators, such as SUMO [4], AIMSUN [5], and MITSIM [6], typically rely on heuristic car-following models like the Intelligent Driver Model (IDM) [7]. However, despite careful calibration of parameters, these simplified, rulebased models often fail to deliver accurate simulations [8] due to the complexity of real-world traffic environments.

We consider the dynamics and the interactions of multiple reinforcement learning optimal execution trading agents interacting with a reactive Agent-Based Model (ABM) of a financial market in event time. The model represents a market ecology with 3-trophic levels represented by: optimal execution learning agents, minimally intelligent liquidity takers, and fast electronic liquidity providers. The optimal execution agent classes include buying and selling agents that can either use a combination of limit orders and market orders, or only trade using market orders. The reward function explicitly balances trade execution slippage against the penalty of not executing the order timeously. This work demonstrates how multiple competing learning agents impact a minimally intelligent market simulation as functions of the number of agents, the size of agents' initial orders, and the state spaces used for learning. We use phase space plots to examine the dynamics of the ABM, when various specifications of learning agents are included. Further, we examine whether the inclusion of optimal execution agents that can learn is able to produce dynamics with the same complexity as empirical data. We find that the inclusion of optimal execution agents changes the stylised facts produced by ABM to conform more with empirical data, and are a necessary inclusion for ABMs investigating market micro-structure. However, including execution agents to chartist-fundamentalist-noise ABMs is insufficient to recover the complexity observed in empirical data.

In this paper, we present a multi-agent deep reinforcement learning (deep RL) framework for network slicing in a dynamic environment with multiple base stations and multiple users. In particular, we propose a novel deep RL framework with multiple actors and centralized critic (MACC) in which actors are implemented as pointer networks to fit the varying dimension of input. We evaluate the performance of the proposed deep RL algorithm via simulations to demonstrate its effectiveness. Subsequently, we develop a deep RL based jammer with limited prior information and limited power budget. The goal of the jammer is to minimize the transmission rates achieved with network slicing and thus degrade the network slicing agents' performance. We design a jammer with both listening and jamming phases and address jamming location optimization as well as jamming channel optimization via deep RL. We evaluate the jammer at the optimized location, generating interference attacks in the optimized set of channels by switching between the jamming phase and listening phase. We show that the proposed jammer can significantly reduce the victims' performance without direct feedback or prior knowledge on the network slicing policies. Finally, we devise a Nash-equilibrium-supervised policy ensemble mixed strategy profile for network slicing (as a defensive measure) and jamming. We evaluate the performance of the proposed policy ensemble algorithm by applying on the network slicing agents and the jammer agent in simulations to show its effectiveness.

The learn-from-observation (LfO) paradigm is a human-inspired mode for a robot to learn to perform a task simply by watching it being performed. LfO can facilitate robot integration on factory floors by minimizing disruption and reducing tedious programming. A key component of the LfO pipeline is a transformation of the depth camera frames to the corresponding task state and action pairs, which are then relayed to learning techniques such as imitation or inverse reinforcement learning for understanding the task parameters. While several existing computer vision models analyze videos for activity recognition, SA-Net specifically targets robotic LfO from RGB-D data. However, SA-Net and many other models analyze frame data captured from a single viewpoint. Their analysis is therefore highly sensitive to occlusions of the observed task, which are frequent in deployments. An obvious way of reducing occlusions is to simultaneously observe the task from multiple viewpoints and synchronously fuse the multiple streams in the model. Toward this, we present multi-view SA-Net, which generalizes the SA-Net model to allow the perception of multiple viewpoints of the task activity, integrate them, and better recognize the state and action in each frame. Performance evaluations on two distinct domains establish that MVSA-Net recognizes the state-action pairs under occlusion more accurately compared to single-view MVSA-Net and other baselines. Our ablation studies further evaluate its performance under different ambient conditions and establish the contribution of the architecture components. As such, MVSA-Net offers a significantly more robust and deployable state-action trajectory generation compared to previous methods.

Peer-to-peer (P2P) energy trading and energy communities have garnered much attention over in recent years due to increasing investments in local energy generation and storage assets. However, the efficiency to be gained from P2P trading, and the structure of local energy markets raise many important challenges. To analyse the efficiency of P2P energy markets, in this work, we consider two different popular approaches to peer-to-peer trading: centralised (through a central market maker/clearing entity) vs. fully decentralised (P2P), and explore the comparative economic benefits of these models. We focus on the metric of Gains from Trade (GT), given optimal P2P trading schedule computed by a schedule optimiser. In both local market models, benefits from trading are realised mainly due to the diversity in consumption behaviour and renewable energy generation between prosumers in an energy community. Both market models will lead to the most promising P2P contracts (the ones with the highest Gains from Trade) to be established first. Yet, we find diversity decreases quickly as more peer-to-peer energy contracts are established and more prosumers join the market, leading to significantly diminishing returns. In this work, we aim to quantify this effect using real-world data from two large-scale smart energy trials in the UK, i.e. the Low Carbon London project and the Thames Valley Vision project. Our experimental study shows that, for both market models, only a small number of P2P contracts, and only a fraction of total prosumers in the community are required to achieve the majority of the maximal potential Gains from Trade. We also study the effect that diversity in consumption profiles has on overall trading potential and dynamics in an energy community.

The safe control of multi-robot swarms is a challenging and active field of research, where common goals include maintaining group cohesion while simultaneously avoiding obstacles and inter-agent collision. Building off our previously developed theory for distributed collaborative safety-critical control for networked dynamic systems, we propose a distributed algorithm for the formation control of robot swarms given individual agent dynamics, induced formation dynamics, and local neighborhood position and velocity information within a defined sensing radius for each agent. Individual safety guarantees for each agent are obtained using rounds of communication between neighbors to restrict unsafe control actions among cooperating agents through safety conditions derived from high-order control barrier functions (CBFs). We provide conditions under which a swarm is guaranteed to achieve collective safety with respect to multiple obstacles using a modified collaborative safety algorithm. We demonstrate the performance of our distributed algorithm via simulation in a simplified physics-based environment.

This paper reports a new hierarchical architecture for modeling autonomous multi-robot systems (MRSs): a non-linear dynamical opinion process is used to model high-level group choice, and multi-objective behavior optimization is used to model individual decisions. Using previously reported theoretical results, we show it is possible to design the behavior of the MRS by the selection of a relatively small set of parameters. The resulting behavior - both collective actions and individual actions - can be understood intuitively. The approach is entirely decentralized and the communication cost scales by the number of group options, not agents. We demonstrated the effectiveness of this approach using a hypothetical `explore-exploit-migrate' scenario in a two hour field demonstration with eight unmanned surface vessels (USVs). The results from our preliminary field experiment show the collective behavior is robust even with time-varying network topology and agent dropouts.

Socially aware robot navigation is gaining popularity with the increase in delivery and assistive robots. The research is further fueled by a need for socially aware navigation skills in autonomous vehicles to move safely and appropriately in spaces shared with humans. Although most of these are ground robots, drones are also entering the field. In this paper, we present a literature survey of the works on socially aware robot navigation in the past 10 years. We propose four different faceted taxonomies to navigate the literature and examine the field from four different perspectives. Through the taxonomic review, we discuss the current research directions and the extending scope of applications in various domains. Further, we put forward a list of current research opportunities and present a discussion on possible future challenges that are likely to emerge in the field.

Due to the great difficulty of thoroughly testing video game software by hand, it is desirable to have AI agents that can automatically explore different game functionalities. A key requirement of such agents is a model of the player actions that the agent can use to both determine the set of possible actions in different game states, as well as perform a chosen action on the game selected by the agent's policy. The typical game engines that are in use today do not offer such a model of actions, leading existing work to either require human effort to manually define the action model or imprecisely guess the possible actions. In our work, we demonstrate how program analysis is an effective solution to this problem by developing a state-of-the-art analysis for the user input handling logic present in games that can automatically model game actions with a discrete action space. Our key insight is that the possible actions of games correspond to the different execution paths that can be taken through the user input handling logic present in the game's code.