Learning-based Management Research Articles

Cooperative Multiagent Deep Reinforcement Learning for Reliable Surveillance via Autonomous Multi-UAV Control

CCTV-based surveillance using unmanned aerial vehicles (UAVs) is considered a key technology for security in smart city environments. This paper creates a case where the UAVs with CCTV-cameras fly over the city area for flexible and reliable surveillance services. UAVs should be deployed to cover a large area while minimize overlapping and shadow areas for a reliable surveillance system. However, the operation of UAVs is subject to high uncertainty, necessitating autonomous recovery systems. This work develops a multi-agent deep reinforcement learning-based management scheme for reliable industry surveillance in smart city applications. The core idea this paper employs is autonomously replenishing the UAV's deficient network requirements with communications. Via intensive simulations, our proposed algorithm outperforms the state-of-the-art algorithms in terms of surveillance coverage, user support capability, and computational costs.

An Energy Management Strategy of Power-Split Hybrid Electric Vehicles Using Reinforcement Learning

With the rapid development of science and technology, the automobile industry is also gradually expanding due to which energy security and ecological security are seriously threatened. This paper was aimed at studying the energy organization strategy of power-split hybrid electric vehicles based on a reinforcement learning algorithm. A power-split hybrid electric vehicle (HEV) combines the advantages of both series and parallel hybrid vehicle architectures by using a planetary gear set to split and combine the power generated by electric machines and a combustion engine. This improves the fuel economy to some extent. However, to increase the fuel economy to a greater extent. This study primarily introduces the hybrid electric vehicle’s structure and presents a reinforcement learning-based management of energy approach for hybrid electric vehicles. It constructs the vehicle power model of HEV and Markov probability transfer model, then designs the energy control strategy based on reinforcement learning, and finally compares it with the energy control strategy based on PID (Proportional-Integral-Derivative). Using MATLAB/Simulink, the cycle conditions of NEDC (New European Driving Cycle) and FTP-75 (Federal Test Procedure) are selected to carry out simulation experiments. The energy management technique suggested in this study, which is based on reinforcement learning, may efficiently enhance the usage rate of automotive gasoline. The replication results show that the fuel consumption per 100 km (kilometers) based on reinforcement learning management strategy is 4.6% and 2.7% lower than the PID management strategy under two working conditions. The economy of fuel of the hybrid electric vehicle is also effectively improved.

The Credibility of International Non-Governmental Organizations (INGOs) and the Oxfam Scandal of 2018

ABSTRACT In 2018, one of the largest international development non-governmental organizations (NGOs) in the world, Oxfam GB, became engulfed in a scandal which quickly spread to other international NGOs (INGOs). The crisis arose from the sexual exploitation, abuse and harassment (SEAH) of the beneficiaries and staff of leading INGOs and caused significant reputational harm to these organizations amid declining public trust and intense political and media scrutiny. The crisis raises significant questions about the credibility of INGOs and the policies necessary to restore public trust. This article reviews the background to the crisis and the responses to it from Oxfam GB & Oxfam International, by other INGOs and by the funders and regulators tasked with overseeing them, focusing on the United Kingdom. It then analyses these actions in the context of an analytical framework proposed in Gourevitch, Lake & Stein (Eds)(2012). It argues that the Oxfam scandal of 2018 marks a fundamental shift in the manner in which INGOs must promote accountability and transparency, based on high-quality, culturally-inclusive, learning-based management.

Machine Learning-Based Management of Electric Vehicles Charging: Towards Highly-Dispersed Fast Chargers

Coordinated charging of electric vehicles (EVs) improves the overall efficiency of the power grid as it avoids distribution system overloads, increases power quality, and decreases voltage fluctuations. Moreover, the coordinated charging supports flattening the load profile. Therefore, an effective coordination technique is crucial for the protection of the distribution grid and its components. The substantial power used through charging EVs has undeniable negative impacts on the power grid. Additionally, with the increasing use of EVs, an effective solution for the coordination of EVs charging, particularly when considering the anticipated proliferation of EV fast chargers, is imminently required. In this paper, different machine learning (ML) approaches are compared for the coordination of EVs charging. The ML models can predict the power to be used in EVs charging stations (EVCS). Due to its ability to use historical data to learn and identify patterns for making future decisions with minimal user intervention, ML has been utilized. ML models used in this paper are (1) Decision Tree (DT), (2) Random Forest (RF), (3) Support Vector Machine (SVM), (4) Naïve Bayes (NB), (5) K-Nearest Neighbors (KNN), (6) Deep Neural Networks (DNN), and (7) Long Short-Term Memory (LSTM). These approaches are chosen as they are classifiers known to have the leading results for multiclass classification problems. The results found shed insight on the importance of the techniques used and their high potential in providing a reliable solution for the coordinated charging of EVs, thus improving the performance of the power grid, and reducing power losses and voltage fluctuations. The use of ML provides a less complex method to coordinate EVs, in comparison with conventional optimization techniques such as quadratic programming, and the use of ML is faster as it requires less computational power. LSTM provided the best results with an accuracy of 95% for predicting the most appropriate power rating (PR) for EVCS, followed by RF, DT, DNN, SVM, KNN, and NB. Additionally, LSTM was also the model with the smallest error rate, at a value of ±0.7%, followed by RF, DT, KNN, SVM, DNN, and NB. The results obtained from the LSTM model were similar to the results obtained from past literature using quadratic programming, with the increased speed and simplicity of ML.

Enhanced Phase-Driven $Q$ -Learning-Based DRM for Multicore Processors

In this paper, we propose a new dynamic reliability management technique for multicore processors using phase-driven ${Q}$ -learning-based method. Our technique considers a wide range of long-term reliability issues and maximizes the throughput of the processor subject to the reliability constraint. We employ ON/OFF switching actions and dynamic voltage and frequency scaling as control knobs (i.e., working modes) to tune the state of cores of the processor. In order to achieve this, our technique detects program phases and adaptively determines the optimal working modes for each phase using the ${Q}$ -learning-based method. By integrating the phase detection into the ${Q}$ -learning-based management, our technique can provide efficient management for the programs with highly diverse phases. We also propose three additional modules to improve the management efficiency of our technique. In order to evaluate our technique, we use it to manage a 3-D CPU with high-diver programs. Several failure mechanisms are considered in this case study. Our proposed technique is compared with two existing ${Q}$ -learning-based techniques. The experimental results demonstrate that when the number of phases is smaller than the number of working modes, our technique can achieve more than $1.36{ \times }$ improvement in performance with 60% memory space savings.

The international legal framework for the management of the global oceans social-ecological system

This paper evaluates the international agreements in place for the protection of the environment and the regulation of human activities taking place in world's oceans and seas. 500 multilateral agreements were reviewed against a framework of reference, grounded on the theoretical approaches of Adaptive Management and Transition Management. According to this framework, oceans complex systems management should: (1) consider the global oceans as a Social-Ecological System (SES); (2) aim to achieve or maintain their ecological resilience; and (3) implement iterative, learning-based management strategies, supported by science-based advice to policy and management. The results show that the present international legal framework for the global oceans does not require countries to adopt an adaptive, complex systems approach for global oceans ecological resilience. Instead, this study supports the perspective of a double fragmentation among international agreements. First, global agreements focus on issue-based objectives for determined human activities, ecological components or anthropogenic pressures. Second, regional agreements have a wider scope, but also a varying level of inclusion of ecological resilience considerations. There is the need to foster the inclusion of such an approach into existing and future international agreements and their implementation, including through soft-law, project-based initiatives at global and regional scales.

The EU legal framework for the management of marine complex social–ecological systems

This paper evaluates the European Union (EU) legal framework for the management of marine complex, adaptive systems. The entire EU legal framework, consisting of 12421 Directives, Regulations and Decisions, is reviewed against a framework of reference, grounded on the theoretical approaches of Adaptive Management and Transition Management. According to this framework, marine complex systems management should: (1) be calibrated at the scale of social–ecological systems; (2) aim to achieve or maintain their ecological resilience; and (3) implement iterative, learning-based management strategies, supported by periodical assessments and monitoring. The results show that the EU legislation does not provide a fully coherent framework for the implementation of a complex systems approach to the management of EU marine social–ecological systems. Although the Marine Strategy Framework Directive (MSFD) is a major step towards this purpose, the present research highlights three major limitations: (1) the limited capacity of the MSFD to support the coordination between Member States sharing the same marine region or sub-region; (2) the insufficient characterisation of marine ecological resilience, in particular in relation to socio-economic elements, ecosystem services, human benefits and cross-scale interactions; and (3) the limited capacity of the MSFD to tackle the fragmentation of the EU legal framework and prioritise complexity and ecological resilience over sectorial approaches.

Adaptive Management: From More Talk to Real Action

The challenges currently facing resource managers are large-scale and complex, and demand new approaches to balance development and conservation goals. One approach that shows considerable promise for addressing these challenges is adaptive management, which by now is broadly seen as a natural, intuitive, and potentially effective way to address decision-making in the face of uncertainties. Yet the concept of adaptive management continues to evolve, and its record of success remains limited. In this article, we present an operational framework for adaptive decision-making, and describe the challenges and opportunities in applying it to real-world problems. We discuss the key elements required for adaptive decision-making, and their integration into an iterative process that highlights and distinguishes technical and social learning. We illustrate the elements and processes of the framework with some successful on-the-ground examples of natural resource management. Finally, we address some of the difficulties in applying learning-based management, and finish with a discussion of future directions and strategic challenges.

Building collaboration and learning in integrated catchment management: the importance of social process and multiple engagement approaches

Success at integrated catchment management (ICM) requires the ongoing participation of different stakeholders in an adaptive and learning-based management process. However, this can be difficult to achieve in practice because many initiatives fail to address the underlying social process aspects required. We review emerging lessons around how to engage stakeholders in ways that support social learning. We focus on the experience of an ICM research programme based in the Motueka catchment in New Zealand and provide a simple framework for distinguishing a range of conversations across different communities of practice. We highlight the need to use multiple engagement approaches to address different constituent needs and opportunities, and to encourage the informal conversations that spring up around these. We then illustrate the range of platforms for dialogue and learning that were used in the programme during 10 years of ICM research. Finally, a number of lessons are described from across the programme to guide research leaders and managers seeking to improve collaboration in other integrated science, management and policy initiatives.