Article
Economics
Thomas W. L. Norman
Summary: There are concerns about reinforcement algorithms learning tacit collusion in oligopolistic pricing and their difficulty in regulation. This study shows the connection between evolutionary game theory and reinforcement learning, demonstrating when the latter's rest points are Bayes-Nash equilibria and deriving a Pigouvian tax system for implementing an unknown socially optimal outcome. Simulation results illustrate the collusion capacity of reinforcement algorithms in a simple setting, but the introduction of optimal taxes leads to a competitive outcome.
JOURNAL OF ECONOMIC BEHAVIOR & ORGANIZATION
(2023)
Article
Computer Science, Hardware & Architecture
Shaohua Cao, Di Liu, Congcong Dai, Chengqi Wang, Yansheng Yang, Weishan Zhang, Danyang Zheng
Summary: With the development of autonomous and intelligent techniques, vehicles are equipped with computation and communication modules to handle on-vehicle computing requests. However, due to limited computation capacities, these requests are offloaded to special devices like roadside units or intelligent vehicles. Two challenges arise in vehicular edge computing networks: accurately determining peak or low hours and effectively offloading requests. This paper investigates computational requests offloading in different vehicular networking scenarios and proposes algorithms based on fuzzy inference and reinforcement learning to address these challenges. Experimental results show significant improvement in resource utilization compared to the benchmark.
Article
Engineering, Multidisciplinary
Jinming Li, Bo Gu, Zhen Qin, Yu Han
Summary: This article proposes a multi-access edge computing system based on vehicle-to-infrastructure communication, and applies a deep reinforcement learning algorithm to task offloading to minimize delay and energy consumption. Experimental results show that the proposed method outperforms existing algorithms.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)
Article
Computer Science, Hardware & Architecture
Alberto Robles-Enciso, Antonio F. Skarmeta
Summary: This paper proposes using reinforcement learning to solve the task assignment problem at the edge layer and extends it with multi-layer reinforcement learning to improve performance. The task assignment process is formulated considering the trade-off between energy consumption and execution time. Simulation results demonstrate the superiority of reinforcement learning solutions over heuristic-based solutions and the performance improvement of the multi-layer approach in high device density scenarios.
Article
Computer Science, Hardware & Architecture
Xiang Ju, Shengchao Su, Chaojie Xu, Haoxuan Wang
Summary: This paper proposes a computation offloading and task scheduling scheme based on pointer network, which can maximize the number of offloading executions of computation tasks and satisfy the priority of computation offloading. The tasks are divided into two types based on whether they can be executed on the vehicle's device, and a two-stage offloading policy is provided. Then, a task offloading decision and scheduling scheme based on pointer network is proposed. Finally, a deep reinforcement learning algorithm is used to train the pointer network for task offloading decision-making and scheduling. The experimental results demonstrate the effectiveness of the proposed scheme.
Article
Computer Science, Information Systems
Xiaoqi Zhang, Hongju Cheng, Zhiyong Yu, Neal N. Xiong
Summary: In this article, a multiresource allocation system for cooperative computing in the Internet of Things based on deep reinforcement learning is proposed. By redefining calculation models and considering practical interference factors, the system efficiently supports complex applications. Experiments have shown that this system has low service latency under resource-constrained conditions, and the improvement is more significant with the increase of network size.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Artificial Intelligence
Weiyao Meng, Rong Qu
Summary: This paper introduces the AutoGCOP framework for the automated design of local search algorithms, optimizing the composition of algorithmic components and utilizing learning models for enhancement. The Markov chain model demonstrates superior performance in learning algorithmic component compositions.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Theory & Methods
Amanda Jayanetti, Saman Halgamuge, Rajkumar Buyya
Summary: The wide-spread adoption of IoT has led to a surge in data generation and processing needs, with edge computing emerging as a complementary solution to cloud computing for IoT applications. This study introduces a novel workflow scheduling framework based on Deep Reinforcement Learning to efficiently handle complex workload scheduling in edge-cloud environments.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2022)
Article
Computer Science, Theory & Methods
Yanqi Gong, Kun Bian, Fei Hao, Yifei Sun, Yulei Wu
Summary: Due to the proliferation of applications such as virtual reality and online games with high real-time requirements, Mobile Edge Computing (MEC) has become a promising computing paradigm that can improve user experience and reduce task offloading latency. However, existing offloading solutions often ignore the important factor of economic cost, making it hard for these solutions to achieve sustainable cloud-edge-end collaborative computation.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2023)
Article
Computer Science, Hardware & Architecture
Hamidreza Mahini, Amir Masoud Rahmani, Seyyedeh Mobarakeh Mousavirad
Summary: The research addresses the IoT task offloading challenge by preparing a set of Python tasks for evaluation, proposing a four-tier architecture for decision-making, formulating the problem as an evolutionary game, and simulating the scheme in MATLAB. The proposed approach contributes to core traffic decreases and has a convergence time of less than 6 seconds for solving problems with 100 tasks.
JOURNAL OF SUPERCOMPUTING
(2021)
Article
Computer Science, Hardware & Architecture
Zahra Aghapour, Saeed Sharifian, Hassan Taheri
Summary: The application of Artificial Intelligence in Internet of Things is increasing, and researchers are searching for methods to overcome the limitations of IoT devices in processing and storing massive computations. Convolutional Neural Network (CNN) processing is a common application for object detection and image classification. By offloading CNN segmented layers to edge cloudlet servers, the total latency and energy consumption of IoT devices can be optimized.
Article
Computer Science, Artificial Intelligence
Zhenhao Wu, Jianbo Gao, Jiashuo Zhang, Yue Li, Qingshan Li, Zhi Guan, Zhong Chen
Summary: Federated learning is a widely researched area due to its ability to break data islands. It brings positive impacts on privacy protection and data cooperation by not uploading client data to a central server. However, some clients are unwilling to participate in federated learning due to resource limitations or potential economic risks. To encourage participation, DFHelper provides collaborative computing services for powerless clients and an incentive mechanism for all roles involved, relieving computing pressure and providing reasonable rewards.
APPLIED INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Neeraj Kumar Rathore, Yunus Khan, Sudesh Kumar, Pawan Singh, Sunita Varma
Summary: This article proposes an evolutionary modern algorithm automated forensic platform based on the concept of Blockchain to address the authenticity issue in cloud computing. The platform utilizes a secure block verification mechanism and optimization algorithms to ensure the collection, storage, and encryption of data. It also utilizes smart contracts and graph neural networks to enable evidence analysis. Experimental results show that the platform achieves good performance in terms of evidence response time, insertion times of cloud evidence, verification time of evidence, etc.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Yue Han, Dusit Niyato, Cyril Leung, Chunyan Miao, Dong In Kim
Summary: This article proposes the concept of coded edge federation (CEF) to address the limited computational resources faced by edge infrastructure providers (EIPs) in supporting CDC tasks in edge networks. The CEF is modeled using evolutionary game theory, and fractional replicator dynamics with power-law fading memory is introduced to support EIPs that are economically aware. Theoretical analysis and numerical experiments confirm the existence and stability of equilibrium in the fractional evolutionary game.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Chemistry, Analytical
Fawzy Habeeb, Tomasz Szydlo, Lukasz Kowalski, Ayman Noor, Dhaval Thakker, Graham Morgan, Rajiv Ranjan
Summary: This paper proposes a reinforcement learning-based dynamic data stream mechanism for time-critical IoT systems in energy-aware IoT devices. The mechanism adjusts the data transport rate based on the available amount of renewable energy resources to ensure reliable data collection while considering the sensor battery lifetime.
Article
Statistics & Probability
Oluwasegun Taiwo Ojo, Antonio Fernandez Anta, Rosa E. Lillo, Carlo Sguera
Summary: This study introduces two new outlier detection methods, "Semifast-MUOD" and "Fast-MUOD", which are based on the existing MUOD method. Evaluation results demonstrate significant improvements in outlier detection and computational time for the proposed methods, with Fast-MUOD particularly suitable for big and dense functional datasets.
ADVANCES IN DATA ANALYSIS AND CLASSIFICATION
(2022)
Article
Multidisciplinary Sciences
Ignacio Tamarit, Angel Sanchez, Jose A. Cuesta
Summary: This paper discusses resource allocation problems in human relationship patterns using a new formalism that allows for continuous values of resource costs. The study shows that there is a characteristic parameter eta, analogous to the ratio of relationships between adjacent circles in the discrete case, which has a value similar to 6. The findings suggest the existence of a universal feature in how humans manage relationships, and that online personal networks mirror offline relationship structures.
SCIENTIFIC REPORTS
(2022)
Article
Multidisciplinary Sciences
Juan Ozaita, Andrea Baronchelli, Angel Sanchez
Summary: Visible markers have minimal impact on promoting cooperation, calling for further research.
SCIENTIFIC REPORTS
(2022)
Article
Computer Science, Information Systems
Mohamed Moulay, Rafael Garcia Leiva, Pablo J. Rojo Maroni, Fernando Diez, Vincenzo Mancuso, Antonio Fernandez Anta
Summary: CIAN is a practical and interpretable ML methodology designed to automate the identification of performance anomalies in cellular networks, using unsupervised learning and combining multiple ML algorithms and data analysis tools. Experimental results show that TTrees implementation of CIAN can automatically identify and accurately classify network anomalies, facilitating precise troubleshooting actions.
COMPUTER COMMUNICATIONS
(2022)
Article
Economics
Pablo Branas-Garza, Diego Jorrat, Antonio M. Espin, Angel Sanchez
Summary: The use of real decision-making incentives in economic experiments remains controversial. This study analyzes data from lab experiments in Spain and field and online experiments in Nigeria and the UK, finding that the use of hypothetical or real payments has minimal impact on the elicitation of short-term and long-term discounting.
EXPERIMENTAL ECONOMICS
(2023)
Article
Multidisciplinary Sciences
Antonio Cabrales, Ryan Kendall, Angel Sanchez
Summary: This study examines policies aimed at reducing negative externalities and explores their differential impact on genders. Through a driving experiment, it is found that slower driving speeds are the safest option, while faster speeds have higher potential payoffs but also higher accident probabilities. The study further reveals that policies utilizing different framing and endogenously determined punishment mechanisms are only effective for female drivers.
Article
Economics
Pablo Branas-Garza, Antonio Cabrales, Guillermo Mateu, Angel Sanchez, Angela Sutan
Summary: This study experimentally examines the impact of pre-play social interactions on negotiations, including conversations, food, and beverages. The results show that none of these interaction components significantly improve negotiation outcomes compared to no interaction. The study also finds no superiority of interaction in terms of trust and reciprocity.
JOURNAL OF BEHAVIORAL AND EXPERIMENTAL ECONOMICS
(2023)
Article
Multidisciplinary Sciences
Miguel Ruiz-Garcia, Juan Ozaita, Maria Pereda, Antonio Alfonso, Pablo Branas-Garza, Jose A. Cuesta, Angel Sanchez
Summary: Networks of social interactions are crucial for civilizations, but the quantitative understanding of them is still limited. This study examines real social networks in 13 schools, involving over 3,000 students and 60,000 relationships. The triadic influence metric is introduced to measure the impact of nearest neighbors, which outperforms personal traits in predicting relationship signs.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2023)
Article
Multidisciplinary Sciences
Miguel A. Gonzalez Casado, Angel Sanchez, Maxi San Miguel
Summary: In this work, the authors assess the role of the adaptation of the interaction network among agents in reaching global coordination and equilibrium selection. They find that the system exhibits fragmentation before reaching full coordination, but coevolution enhances the selection of payoff-dominant equilibrium in a coordination game with risk. Moreover, there is an intermediate range of plasticity values where the system fully coordinates on a single component network.
SCIENTIFIC REPORTS
(2023)
Article
Statistics & Probability
Oluwasegun Taiwo Ojo, Antonio Fernandez Anta, Marc G. G. Genton, Rosa E. E. Lillo
Summary: This article introduces the definitions and properties of the fast massive unsupervised outlier detection (FastMUOD) indices in functional data. FastMUOD detects outliers by calculating amplitude, magnitude, and shape indices for each curve to target the corresponding types of outliers. Several methods adapting FastMUOD to outlier detection in multivariate functional data are proposed, including applying FastMUOD on the components of the multivariate data and using random projections. Furthermore, these techniques are tested on various simulated and real multivariate functional datasets. Compared with the state of the art in multivariate functional OD, the use of random projections showed the most effective results with similar, and in some cases improved, OD performance. Based on the proportion of random projections that flag each multivariate function as an outlier, a new graphical tool called the magnitude-shape-amplitude (MSA) plot is proposed for visualizing the magnitude, shape, and amplitude outlyingness of multivariate functional data.
Article
Humanities, Multidisciplinary
Denis Tverskoi, Andrea Guido, Giulia Andrighetto, Angel Sanchez, Sergey Gavrilets
Summary: In social interactions, human decision-making, attitudes, and beliefs about others are influenced by factors such as cognitive processes, social influences, and cost-benefit considerations. This study aims to understand these dynamics by using mathematical modeling and an online behavioral experiment. The results show that personal norms and conformity with peers have the largest impact on decision-making, while material benefits and normative expectations have smaller effects. Messaging can change these dynamics by decreasing personal norms and increasing conformity. The study highlights the importance of understanding the dynamics of personal beliefs and beliefs about others in social behavior.
HUMANITIES & SOCIAL SCIENCES COMMUNICATIONS
(2023)
Article
Multidisciplinary Sciences
Rosendo Castanon, Fco. Alberto Campos, Jose Villar, Angel Sanchez
Summary: This paper proposes an agent-based model that utilizes the Bush-Mosteller reinforcement learning algorithm to investigate the role of social expectations in the emergence of altruistic behavior. The model is compared with experimental results, and the findings suggest that a stimuli model based on empirical and normative expectations has potential in understanding decision-making processes in contexts where cooperative behavior is relevant.
SCIENTIFIC REPORTS
(2023)
Article
Multidisciplinary Sciences
Pablo Lozano, Alberto Antonioni, Angel Sanchez
Summary: This study uses online human experiments to investigate the effects of hierarchical structures on cooperation and conflict within groups. The results show that cooperation can be maintained through dyadic conflicts that punish noncooperators, leading to stable hierarchical groups with high levels of cooperation. These findings align with earlier theoretical models and provide insights into the relationship between hierarchy, cooperation, and conflict.
Article
Multidisciplinary Sciences
Pablo Lozano, Alberto Antonioni, Angel Sanchez
Summary: This study used online human experiments to investigate the effects of hierarchy on cooperation and conflict within groups. The results showed that cooperation can be maintained through participating in conflicts, leading to stable hierarchical groups with high levels of cooperation.
Article
Multidisciplinary Sciences
Jose Luis Molina, Juan Ozaita, Ignacio Tamarit, Angel Sanchez, Christopher McCarty, H. Russell Bernard
Summary: Culture and social structure are intertwined phenomena observable in personal networks, and can be predicted by specific combinations of personal network structural measures such as closeness, clustering, betweenness, and average degree. The findings support the theory of Grid/Group that emphasizes the interdependence of social structural and cultural features.
