Cooperative multi-agent decision-making is a ubiquitous problem with many real-world applications. In many practical applications, it is desirable to design a multi-agent team with a heterogeneous composition where the agents can have different capabilities and levels of risk tolerance to address diverse requirements. While heterogeneity in multi-agent teams offers benefits, new challenges arise including how to find optimal heterogeneous team compositions and how to dynamically distribute tasks among agents in complex operations. In this work, we develop an artificial intelligence framework for multi-agent heterogeneous teams to dynamically learn task distributions among agents through reinforcement learning. The framework extends Decentralized Partially Observable Markov Decision Processes (Dec-POMDP) to be compatible to model various types of heterogeneity. We demonstrate our approach with a benchmark problem on a disaster relief scenario. The effect of heterogeneity and risk aversion in agent capabilities and decision-making strategies on the performance of multi-agent teams in uncertain environments is analyzed. Results show that a well-designed heterogeneous team outperforms its homogeneous counterpart and possesses higher adaptivity in uncertain environments.

Truck platooning enabled by connected automated vehicle (CAV) technology has been demonstrated to effectively reduce fuel consumption for trucks in a platoon. However, given the limited number of trucks in the traffic stream, it remains questionable how great an energy saving it may yield for a practical freight system if we only rely on ad-hoc platooning. Assuming the presence of a central platooning coordinator, this paper is offered to substantiate truck platooning benefits in fuel economy produced by exploiting platooning opportunities arising from the United States’ domestic truck demands on its highway freight network. An integer programming model is utilized to schedule trucks’ itineraries to facilitate the formation of platoons at platoonable locations to maximize energy savings. A simplification of the real freight network and an approximation algorithm are used to solve the model efficiently. By analyzing the numerical results obtained, this study quantifies the importance of scheduled platooning in improving trucks’ fuel economy. Furthermore, the allowable platoon size, schedule flexibility, and fuel efficiency all play a crucial role in energy savings. Specifically, by assuming that following vehicles in a platoon obtain a 10% energy reduction, an average energy reduction of 8.48% per truck can be achieved for the overall network if the maximum platoon size is seven, and the schedule flexibility is 30 min. The cost–benefit analysis provided at the end suggests that the energy-saving benefits can offset the investment cost in truck platooning technology.

In operations of multi-agent teams ranging from homogeneous robot swarms to heterogeneous human-autonomy teams, unexpected events might occur. While efficiency of operation for multi-agent task allocation problems is the primary objective, it is essential that the decision-making framework is intelligent enough to manage unexpected task load with limited resources. Otherwise, operation effectiveness would drastically plummet with overloaded agents facing unforeseen risks. In this work, we present a decision-making framework for multiagent teams to learn task allocation with the consideration of load management through decentralized reinforcement learning, where idling is encouraged and unnecessary resource usage is avoided. We illustrate the effect of load management on team performance and explore agent behaviors in example scenarios. Furthermore, a measure of agent importance in collaboration is developed to infer team resilience when facing handling potential overload situations.

Many natural, engineered and social systems can be represented using the framework of a layered network, where each layer captures a different type of interaction between the same set of nodes. The study of such multiplex networks is a vibrant area of research. Yet, understanding how to quantify the correlations present between pairs of layers, and more so present in their co-evolution, is lacking. Such methods would enable us to address fundamental questions involving issues such as function, redundancy, and potential disruptions. Here, we show first how the edge set of a multiplex network can be used to construct an estimator of a joint probability distribution describing edge existence over all layers. We then adapt an information-theoretic measure of general correlation called the conditional mutual information, which uses the estimated joint probability distribution, to quantify the pairwise correlations present between layers. The pairwise comparisons can also be temporal, allowing us to identify if knowledge of a certain layer can provide additional information about the evolution of another layer. We analyse datasets from three distinct domains—economic, political, and airline networks—to demonstrate how pairwise correlation in structure and dynamical evolution between layers can be identified and show that anomalies can serve as potential indicators of major events such as shocks.

Many natural, engineered, and social systems can be represented using the framework of a layered network, where each layer captures a different type of interaction between the same set of nodes.The study of such multiplex networks is a vibrant area of research. Yet, understanding how to quantify the correlations present between pairs of layers, and more so present in their co-evolution, is lacking. Such methods would enable us to address fundamental questions involving issues such as function, redundancy and potential disruptions. Here we show first how the edge-set of a multiplex network can be used to construct an estimator of a joint probability distribution describing edge existence over all layers. We then adapt an information-theoretic measure of general correlation called the conditional mutual information, which uses the estimated joint probability distribution, to quantify the pair-wise correlations present between layers. The pair-wise comparisons can also be temporal, allowing us to identify if knowledge of a certain layer can provide additional information about the evolution of another layer. We analyze datasets from three distinct domains-economic, political, and airline networks-to demonstrate how pair-wise correlation in structure and dynamical evolution between layers can be identified and show that anomalies can serve as potential indicators of major events such as shocks.