Global cooperation on climate change mitigation is essential to limit temperature increases while supporting long-term, equitable economic growth and sustainable development. Achieving such cooperation among diverse regions, each with different incentives, in a dynamic environment shaped by complex geopolitical and economic factors, without a central authority, is a profoundly challenging game-theoretic problem. This article introduces RICE-N, a multi-region integrated assessment model that simulates the global climate, economy, and climate negotiations and agreements. RICE-N uses multi-agent reinforcement learning (MARL) to encourage agents to develop strategic behaviors based on the environmental dynamics and the actions of the others. We present two negotiation protocols: (1) Bilateral Negotiation, an exemplary protocol and (2) Basic Club, inspired from Climate Clubs and the carbon border adjustment mechanism (Nordhaus, 2015; Comissions, 2022). We compare their impact against a no-negotiation baseline with various mitigation strategies, showing that both protocols significantly reduce temperature growth at the cost of a minor drop in production while ensuring a more equitable distribution of the emission reduction costs.

WarpDrive is a flexible, lightweight, and easy-to-use open-source framework for end-to-end deep multi-agent reinforcement learning (MARL) on a Graphics Processing Unit (GPU), available at https://github.com/salesforce/warp-drive. It addresses key system bottlenecks when applying MARL to complex environments with high-dimensional state, observation, or action spaces. For example, WarpDrive eliminates data copying between the CPU and GPU and runs thousands of simulations and agents in parallel. It also enables distributed training on multiple GPUs and scales to millions of agents. In all, WarpDrive enables orders-of-magnitude faster MARL compared to common CPU-GPU implementations. For example, WarpDrive yields 2.9 million environment steps/second with 2000 environments and 1000 agents (at least 100× faster than a CPU version) in a 2d-Tag simulation. It is user-friendly: e.g., it provides a lightweight, extendable Python interface and flexible environment wrappers. It is also compatible with PyTorch. In all, WarpDrive offers a platform to significantly accelerate reinforcement learning research and development.

We propose and experimentally demonstrate a novel approach to improve the packet delivery efficiency on a vulnerable downlink (e.g., from a transmitter to a far-away receiver) using superposition coding, a multiuser transmission scheme that forgoes orthogonal transmission and deliberately introduces interference among signals at the transmitter. On a software radio platform that uses off-the-shelf point-to-point channel codes, we show that a transmitter serving multiple links can use simple two-user superposition codes to dramatically improve (compared to time division multiplexing) the packet delivery efficiency on its most vulnerable links. Interestingly, our results suggest that superposing signals of far-away users on to those of high-traffic users yields the maximum benefits - implying that the degrees-of-freedom gain in doing so can more than compensate for the increased interference from signal superposition.

Characterizing the performance of ad hoc networks is one of the most intricate open challenges; conventional ideas based on information-theoretic techniques and inequalities have not yet been able to successfully tackle this problem in its generality. Motivated thus, we promote the totally asymmetric simple exclusion process (TASEP), a particle flow model in statistical mechanics, as a useful analytical tool to study ad hoc networks with random access. Employing the TASEP framework, we first investigate the average end-to-end delay and throughput performance of a linear multihop flow of packets. Additionally, we analytically derive the distribution of delays incurred by packets at each node, as well as the joint distributions of the delays across adjacent hops along the flow. We then consider more complex wireless network models comprising intersecting flows, and propose the partial mean-field approximation (PMFA), a method that helps tightly approximate the throughput performance of the system. We finally demonstrate via a simple example that the PMFA procedure is quite general in that it may be used to accurately evaluate the performance of ad hoc networks with arbitrary topologies.

We design and implement a software-radio system for Superposition Coding (SC), a multiuser transmission scheme that deliberately introduces interference among user signals at the transmitter, using a library of off-the-shelf point-to-point channel codes. We experimentally determine the set of rate-pairs achieved by this transmission scheme under a packet-error constraint. Our results suggest that SC can provide substantial gains in spectral efficiencies over those achieved by orthogonal schemes such as Time Division Multiplexing. Our findings also question the practical utility of the Gaussian approximation for the inter-user interference in Superposition-Coded systems.

Superposition coding is a well-known capacity-achieving coding scheme for stochastically degraded broadcast channels. Although well-studied in theory, it is important to understand issues that arise when implementing this scheme in a practical setting. In this paper, we present a software-radio based design of a superposition coding system on the GNU Radio platform with the Universal Software Radio Peripheral acting as the transceiver frontend. We also study the packet error performance and discuss some issues that arise in its implementation.

We consider a multihop wireless line network with a single unidirectional data flow and show that by limiting the buffer sizes at the relay nodes to unity, the flow of traffic in the system can be efficiently regulated in a completely distributed fashion. Upon exerting this simple transmission policy, we find that the transport of packets in the wireless network is analogous to the flow of particles in the totally asymmetric simple exclusion process (TASEP). Using existing results from statistical mechanics, we characterize the end-to-end delay and throughput performance of multihop wireless line networks for two different channel access schemes. Additionally, we apply our findings towards the design of long networks. This paper also aims at promoting the TASEP as a powerful tool for analyzing the performance of ad hoc networks.

Opportunistic forwarding protocols take advantage of contact opportunities to route data in intermittently connected environments. In these environments, a fully connected path between the source and destination may not always exist and the contact schedules of all the nodes are not known in advance. Hence, one of the key challenges for a node is to make effective forwarding decisions using only a limited knowledge of the contact behavior of the nodes in the network. Based on an analysis of human mobility traces that we collected from our office environment, we introduce a new link metric, conditional residual time, that accurately estimates the time remaining for a pair of nodes to meet using only the local knowledge of their past contacts. We then propose a distributed protocol called CREST, that uses the conditional residual time to opportunistically forward messages between pairs of nodes. Experimental results show that CREST has a lower end-to-end delay compared to protocols that depend on future contact schedules and global knowledge of the contact behavior across the network. Furthermore, by disseminating only a few additional copies of the message, the delivery ratio of CREST improves significantly and is comparable to that of the flooding protocol.

The optimality of transmit beamforming for a multiple antenna system with partial/limited feedback is investigated and a single general necessary and sufficient condition for beamforming to achieve ergodic capacity is derived. The condition obtained is universal - applicable to all partial/limited feedback scenarios in all ergodic fading channel distributions regardless of the number of transmit/receive antennas or transmit power. Using the universal condition we explore the optimality of beamforming for the quantized mean feedback scheme, which unifies previous results for the separate cases of mean feedback and quantized feedback. Numerical results are provided to complement the analysis

The simple, yet powerful concept of an ldquoerristorrdquo and its ldquoerristancerdquo has recently been introduced for ad hoc networks and applied to scenarios such as retransmission (time diversity), path diversity, or a combination thereof. We extend this formalism to the case of spatial diversity, realized by employing multiple antennas at each node. Based on this framework, one can efficiently analyze and design Rayleigh-faded MIMO ad hoc networks that employ selection combining. The mathematically tractable definition of the erristor term greatly simplifies the study of a multiple-antenna network and helps solve problems based on end-to-end reliability or resource allocation easily, which we illustrate in an example. Moreover, this technique demonstrates the superiority in performance of MIMO over single-antenna routing schemes, particularly at high SNR.