Understanding the factors that affect Key Performance Indicators (KPIs) and how they affect them is frequently important in sectors where data and data science are crucial. Machine learning is utilized to model and predict pertinent KPIs in order to do this. Interpretability is important, nevertheless, in order to fully comprehend how the model generates its predictions. It enables users to pinpoint which traits have aided the model’s ability to learn and comprehend the data. A practical approach for evaluating the contribution of input attributes to model learning has evolved in the form of SHAP (SHapley Additive exPlanations offer an index for evaluating the influence of each feature on the forecasts made by the model. In this paper, it is demonstrated that the contribution of features to model learning may be precisely estimated when utilizing SHAP values with decision tree-based models, which are frequently used to represent tabular data.
In the realm of vehicular networks, ensuring top-notch services for high-speed mobile vehicles remains a formidable challenge. This study squarely addresses the spectrum-sharing predicament within the Internet of Vehicles (IoV) by scrutinizing the spectrum shared by V2I links serving multiple V2V links. The swiftness of vehicular movement ushers in challenges for centralized resource management, necessitating an innovative approach. To surmount this obstacle, the resource allocation is casted as a Multi-agent Deep Reinforcement Learning (RL) and a distributed Multi-Agent Deep Deterministic Policy Algorithm (MADDPG) algorithm is introduced. Each agent autonomously interacts with the vehicular network, assessing its local state and receiving a shared reward. Through collaboration, the Critic network refines power control for each agent. The reward function and training mechanism empower the multi-agent algorithm to attain distributed resource allocation, markedly enhancing both V2I link capacity and V2V transmission rates. This pioneering approach optimizes spectrum utilization, ushering in superior service quality for high-speed mobile vehicles in vehicular networks.
1 January 2020 in Journal of Industrial and Management Optimization
In this paper, we mathematically associate Crypto Cloud Computing, that has become an emerging research area, with Cooperative Game Theory in the presence of uncertainty. In the sequel, we retrieve data from the database of Amazon Web Service. The joint view upon Crypto Cloud Computing, Cooperative Game Theory and Uncertainty management is a novel approach. For this purpose, we construct a cooperative interval game model and apply this model to Social Networks. Then, we suggest some interval solutions related with the model by proposing a novel elliptic curve public key encryption scheme over finite fields having the property of semantic security. The paper ends with concluding words and an outlook to future studies.
Purpose: The purpose of this study is to propose a decentralized multi-party cross-trading scheme based on a certificate transaction mechanism for the transaction of excess consumption certificates (ECCs) of renewable energy. The aim is to address the problems associated with the existing centralized transaction mode and to promote the development of the green electricity industry.Design/methodology/approach: The proposed scheme involves calculating the quotation difference for the same type of certificate transaction based on the quotations of all users of both buyers and sellers. The transaction volume is then determined based on the order of quotation difference from large to small, and the total interests of cooperation are calculated. The nucleolus method is adopted to allocate the total interests to each member of the alliance and calculate the final transaction price. The blockchain technology is used for the transaction to achieve accurate traceability and efficient supervision, and a corresponding smart contract is designed and simulated in the Ethereum consortium chain.Findings: The results of the simulation show the rationality and effectiveness of the proposed scheme. The decentralized multi-party cross-trading scheme can overcome the problems associated with the existing centralized transaction mode, such as low transaction efficiency, difficulty in obtaining the optimal transaction strategy and efficient supervision. The proposed scheme can promote the development of the green electricity industry by stimulating users' demand potential for green electricity.Originality/value: The proposed scheme is original in its use of a certificate transaction mechanism to facilitate the trading of ECCs of renewable energy. The scheme adopts a decentralized multi-party cross-trading approach that overcomes the problems associated with the existing centralized transaction mode. The use of the nucleolus method for the allocation of total interests to each member of the alliance is also original. Finally, the use of blockchain technology for accurate traceability and efficient supervision of the transaction is an original contribution to the field.