This study examines the scientific production focused on the Maximum Principle between 1962 and 2021. Results indicate a consistent increase in the absolute number of publications over time. In relative terms, there is a resurgence of interest in this research field after the period between 2004 and 2009. Overall, these findings support the idea of strategic complementarity between the Maximum Principle and optimal control. However, there is a notable exception during the period 2010–2015, characterised by a decline in scientific production focused on the Maximum Principle and a simultaneous increase in focus on optimal control. Academic journals that play a role in promoting this research field tend to have high impact factors and interesting cite scores. Using a modified Boston Consulting Group matrix, the results reveal the persistence of two researchers labelled as stars and three as cash cows. A multiple linear regression analysis confirms that reputation significantly influences the clustering trends. A critical discussion is provided to highlight the dichotomy between popularity and effective contributions in this research field.
We investigate the convergence properties of policy iteration and value iteration algorithms in reinforcement learning by leveraging fixed-point theory, with a focus on mappings that exhibit weak contractive behavior. Unlike traditional studies that rely on strong contraction properties, such as those defined by the Banach contraction principle, we consider a more general class of mappings that includes weak contractions. Employing Zamfirscu’s fixed-point theorem, we establish sufficient conditions for norm convergence in infinite-dimensional policy spaces under broad assumptions. Our approach extends the applicability of these algorithms to feedback control problems in reinforcement learning, where standard contraction conditions may not hold. Through illustrative examples, we demonstrate that this framework encompasses a wider range of operators, offering new insights into the robustness and flexibility of iterative methods in dynamic programming.
This study presents classification and prediction exercises to evaluate the future behavior of nitrogen dioxide in a critical air quality zone located in Portugal using a dataset, the time span of which covers the period between 1 September 2021 and 23 July 2022. Three main results substantiate the importance of this research. First, the classification analysis corroborates the idea of a neutrality principle of road traffic on the target since the respective coefficient is significant, but quantitatively close to zero. This result, which may be the first sign of a paradigm shift regarding the adoption of electric vehicles in addition to reflect the success of previously implemented measures in the city of Lisbon, is reinforced by evidence that the carbon monoxide emitted mostly by diesel vehicles exhibits a significant, negative and permanent effect on satisfying the hourly limit value associated with the target. Second, robustness checks confirm that the period between 8 h and 16 h is particularly remarkable for influencing the target. Finally, the predictive exercise demonstrates that the internationally patented Variable Split Convolutional Attention model has the best predictive performance among several deep learning neural network alternatives. Results indicate that the concentration of nitrogen dioxide is expected to be volatile and only a redundant downward trend is likely to be observed. Therefore, in terms of policy recommendations, additional measures to avoid exceeding the legal nitrogen dioxide ceiling at the local level should be focused on reducing carbon monoxide emissions, rather than just being concerned about halting the intensity of road traffic.
This video shows a brief overview over Neptus, a command and control infrastructure for heterogeneous teams of autonomous vehicles. Having different types of vehicles at our laboratory and from our partners, there was an increasing need to create a common infrastructure to all these systems. Additionally, a tool to support the entire mission life cycle (planning, execution, review and dissemination) was lacking. Neptus was created to provide vehicle independence and seamless inter-systems communications. Currently, Neptus has been already tested with remotely operated vehicles, autonomous underwater vehicles, unmanned air vehicles, autonomous surface vehicles and wireless sensor networks. Some of these systems were operated simultaneously by various operating consoles that were sharing the same communication infrastructure. The received data was being relayed to a Web server that allowed for the real-time mission following by using a common Web browser.
To advance our understanding of the magnetohydrodynamic (MHD) processes in liquid metals, in this paper we propose an approach combining the classical methods in the dynamo theory based on numerical simulations of the partial differential equations governing the evolution of the magnetic field with the statistical methods. In this study, we intend to answer the following ``optimization'' question: Can we find a statistical explanation what makes a flow to favour magnetic field generation in the linear regime (i.e. the kinematic dynamo is considered), where the Lorenz force is neglected? The flow is assumed to be steady and incompressible, and the magnetic field generation is governed by the magnetic induction equation. The behaviour of its solution is determined by the dominant (i.e. with the largest real part) eigenvalue of the magnetic induction operator. Considering an ensemble of 2193 randomly generated flows, we solved the kinematic dynamo problem and performed an attempt to find a correlation between the dominant eigenvalue and the standard quantities used in hydrodynamics -- vorticity and kinetic helicity. We have found that there is no visible relation between the property of the flow to be a kinematic dynamo and these quantities. This enables us to conclude that the problem requires a more elaborated approach to ``recognize'' if the flow is a dynamo or not; we plan to solve it using contemporary data-driven approach based on deep neural networks.