Multiple Performance Indicators Research Articles

Machine learning-based multi-performance prediction and analysis of Earth-Air Heat Exchanger

The Earth-Air Heat Exchanger (EAHE) plays a crucial role in global renewable energy utilization. Despite various EAHE models, a machine learning-based multi-performance prediction model is lacking. This study introduces a framework based on diverse machine learning models for precise prediction and comprehensive analysis of EAHE's multiple performance indicators. Initially, the theoretical EAHE model is established and validated. Six feature variables and four target variables are input to generate a dataset for machine learning. Six machine learning models are then selected for training, optimization, and validation. Subsequently, model interpretation and analysis are conducted using SHapley additive exPlanations and Partial dependence plots. Results indicate the optimized-extreme gradient boosting model performs the best, with a determination coefficient of 0.98, root mean squared error of 22.98, and mean absolute error of 8.88, identified as the top multi-performance prediction model for EAHE. Notably, pipe diameter, air velocity, and ground type significantly influence EAHE performance. Complex correlations and mutual exclusions exist among target variables, with a strong negative correlation (−0.59) between the coefficient of performance and payback time, highlighting the challenge of simultaneous improvement. This study presents a valuable framework for predicting and analyzing EAHE performance, crucial for promoting its widespread application and enhancing renewable energy utilization.

Applying assessment methods to diversified farming systems: Simple adjustment or complete overhaul?

In response to criticism of current specialized and input-intensive agriculture, diversified farming systems (DFS) have increased in popularity. The most advanced forms of DFS include perennial components (e.g., perennial crops, trees), functionally different plants and livestock, and landscape heterogeneity. These characteristics make it more difficult to assess DFS in agricultural sciences, which have focused to date mainly on low-diversity systems. We provide an overview of the main methodological issues involved in assessing DFS, and their potential solutions. We identify five key methodological challenges for assessing DFS: i) combining conceptual frameworks from different disciplines to develop a holistic view of DFS multi-performance; ii) characterizing DFS structure from a functional perspective to define the relevant spatial extent to assess them; iii) considering DFS ontogeny i.e. the time required for DFS to establish, reach maturity and achieve a dynamic steady state when assessing their performance; iv) aggregating and mapping multiple performance indicators to integrate the range of products and services generated by DFS and to reveal potential synergies and antagonisms; v) defining and characterizing the reference systems used to compare DFS, as systems might differ in history, composition, management, soil-climate, and sociotechnical conditions. These challenges underline the need to overhaul assessments in agricultural sciences (e.g., agronomy, animal science, economics, ecology) to identify whether, and under what conditions, DFS are relevant and sustainable options for farmers. They also come with available options for researchers to proceed with this overhaul.

Deep reinforcement learning for adaptive frequency control of island microgrid considering control performance and economy

To achieve frequency stability and economic efficiency in isolated microgrids, grid operators face a trade-off between multiple performance indicators. This paper introduces a data-driven adaptive load frequency control (DD-ALFC) approach, where the load frequency controller is modeled as an agent that can balance different objectives autonomously. The paper also proposes a priority replay soft actor critic (PR-SAC) algorithm to implement the DD-ALFC method. The PR-SAC algorithm enhances the policy randomness by using entropy regularization and maximization, and improves the learning adaptability and generalization by using priority experience replay. The proposed DD-ALFC method based on the PR-SAC algorithm can achieve higher adaptability and robustness in complex microgrid environments with multiple performance indicators, and improve both the frequency control and the economic efficiency. The paper validates the effectiveness of the proposed method in the Zhuzhou Island microgrid.

A Security Analysis Model for IoT-ecosystem Using Machine Learning- (ML) Approach

Introduction:: The attacks on IoT systems are increasing as the devices and communication networks are progressively integrated. If no attacks are found in IoT for a long time, it will affect the availability of services that can result in data leaks and can create a significant impact on the associated costs and quality of services. Therefore, the attacks and security vulnerability in the IoT ecosystem must be detected to provide robust security and defensive mechanisms for real-time applications. Method:: This paper proposes an analytical design of an intelligent attack detection framework using multiple machine learning techniques to provide cost-effective and efficient security analysis services in the IoT ecosystem. Result:: The performance validation of the proposed framework is carried out by multiple performance indicators. Conclusion:: The simulation outcome exhibits the effectiveness of the proposed system in terms of accuracy and F1-score for the detection of various types of attacking scenarios.

Calibration of Heterogeneous Treatment Effects in Randomized Experiments

Machine learning is commonly used to estimate the heterogeneous treatment effects (HTEs) in randomized experiments. Using large-scale randomized experiments on Facebook and Criteo platforms, we observe substantial discrepancies between machine learning-based treatment effect estimates and difference-in-means estimates directly from the randomized experiment. This paper provides a two-step framework for practitioners and researchers to diagnose and rectify this discrepancy. We first introduce a diagnostic tool to assess whether bias exists in the model-based estimates from machine learning. If bias exists, we then offer a model-agnostic method to calibrate any HTE estimates to known, unbiased, subgroup difference-in-means estimates, ensuring that the sign and magnitude of the subgroup estimates approximate the model-free benchmarks. This calibration method requires no additional data and can be scaled for large data sets. To highlight potential sources of bias, we theoretically show that this bias can result from regularization, and further use synthetic simulation to show biases result from misspecification and high-dimensional features. We demonstrate the efficacy of our calibration method using extensive synthetic simulations and two real-world randomized experiments. We further demonstrate the practical value of this calibration in three typical policy-making settings: a prescriptive, budget-constrained optimization framework; a setting seeking to maximize multiple performance indicators; and a multitreatment uplift modeling setting.

A dual-population multiobjective co-evolutionary matching ensemble learning for product multi-indicator prediction in continuous annealing

Achieving simultaneous and precise prediction of multiple product performance indicators is crucial to ensure stable production in the continuous annealing process. However, the majority of existing quality prediction modeling methods concentrate on developing single-output regression models that are limited to single-indicator prediction scenarios. When dealing with the multi-indicator prediction scenarios, these methods usually require the construction of a separate model for each performance indicator, which leads to more computational and storage resources and more complex maintenance in practical applications. To address this issue, a dual-population multiobjective co-evolutionary matching ensemble learning method called DP-CEMEL is developed. It utilizes dual-population co-optimization to enable the joint learning of the algorithm on input–output relationships and inter-indicator correlations. More specifically, a two-stage learning based individual encoding and decoding method and a performance boosting degree (PBD) based individual evaluation strategy are designed to achieve co-optimization of the dual populations to evolve a set of individuals (i.e., base learners) with good accuracy and low complexity. Then, a PBD based matching ensemble method is applied to get the final multi-indicator prediction model by the optimal matching and ensemble of the evolved individuals. Experimental results on both benchmark data and practical strip production data demonstrate that the proposed DP-CEMEL attains competitive or better performance compared to other state-of-the-art learning methods.

Design of S-box multi-objective optimization algorithm based on combined chaotic system

S-box is the only nonlinear cryptographic component that determines the security of the block cipher. The higher the security of the S-box, the higher the security of the block cipher. Therefore, this paper proposes an S-box multi-objective optimization algorithm based on the combined chaotic system. Firstly, designing an integrated chaotic system based on a fractional index and its dynamic behavior is studied; it shows incredibly high-performance stability and chaotic parameter range coverage in the entire parameter space. A novel chaotic S-box construction algorithm is proposed based on the combined chaotic system. It introduces a linear congruential pseudo-random number generator to extend the sequence period and scramble the chaotic S-box through Henon mapping to improve the nonlinearity of the s-box. Finally, introducing a teaching and learning multi-objective optimization model and the performance evaluation criteria of the S-box are incorporated into the design of the objective function; this design enables the resulting S-box to perform well under multiple performance indicators, and then the approximate optimal S-box in the population is obtained. Through the performance test of the approximate optimal S-box, the comparative analysis shows that the S-box has good cryptographic performance, can resist linear analysis and differential analysis, and has a good application prospect in lightweight cipher.

Cam Profile Design of Low-Speed Marine Diesel Engine Based on the Grey Theory

The camshaft profile is one of the important parameters of the engine, the reasonableness or otherwise of its design directly influences the engine’s power output, fuel consumption, emissions, noise and reliability. However, most existing cam profile optimization designs are based on the basic software operation, profile parameter selection and tuning lack a scientific theoretical basis. In order to address this issue, this study establishes a single valve kinematic model of the gas delivery mechanism of diesel engines by using the AVL-EXCITE Timing Drive, and performs a model-based multinomial kinematic acceleration cam profile design. Subsequently, an orthogonal design method is employed to obtain the parameter combinations that can satisfy the dynamic requirements. The research investigates the influence of cam profile design parameters (C 4, p, q, r, and s) on valve fullness, maximum jerk, cam contact stress, and valve seating force through a systematic study. Lastly, based on the grey correlation theory, a thorough analysis of the performance indices of cams with different profile parameter combinations is conducted. The optimization problem of the profile parameters with multiple performance indicators is transformed into a single-objective grey relational optimization problem, resulting in the optimal cam profile parameter combination of C4 =0.1, p=14, q=20, r=28 and s=36. The optimal method proposed in the paper is straightforward to implement and is applicable to the optimum design of cam profiles of the engine, which is of great importance for the optimal design of cam fabrication processes.

Performance of innovative cropping systems diversified with oilseeds and protein crops: identification and resolution of methodological issues, using the Syppre experimental network as a case study

Agroecological transition requires that innovative and diversified cropping systems be developed. Conducting system experiments is an approach well-suited to the analysis of performance of cropping systems when subjected to soil, weather and biotic stresses. Conducting system experiments nevertheless gives rise to methodological challenges. Using the Syppre network of experiments, consisting of five sites in France, we present an original case study that provides valuable methodological and agronomic lessons on system experiments. The innovative cropping systems tested there are based on crop diversification (including oilseeds and protein crops), as well as flexible tillage, technical innovations and optimized crop management. From a methodological standpoint, we show that (i) mixed models are adapted to a range of experimental questions and constraints; (ii) multifactorial analysis enables the characterization of relationships between performance indicators; (iii) a multisite experimental network is an efficient approach not only for answering agronomic questions, but also for addressing methodological issues. From an agronomic standpoint, we showed that reconciling multiple indicators of performance is still challenging. Overall, innovative and diversified systems improved the performance of input utilization and environmental impacts, but with lower productivity and profitability. Introducing legume crops is a promising strategy because this contributes significantly to reductions in mineral N fertilizer use, energy consumption and greenhouse gas emissions, without major trade-offs against other performance indicators. Finally, we showed that the nature of the production situation had a major influence on the performance profile. This led us to be cautious in making overall analyses especially with regard to general conclusions.

A Review of Three Methods of Artificial Intelligence in Smart Grid Cyber Security: Machine Learning, Reinforcement Learning, Ensemble Methods

Renewable energy is gradually replacing traditional fossil fuels. The change of power generation energy structure brings new challenges to the traditional power grid. Through the efficient bidirectional movement of electricity and information, smart grids might include renewable energy. For the complex informational and financial operations required by smart grid, communication systems are crucial, but they also make smart grid vulnerable to numerous cyber attacks. Smart grid cyber security has been widely concerned. The purpose of this paper is to explore the use of artificial intelligence technology in smart grid cyber security. Three methods in the field of artificial intelligence are highlighted: Machine Learning, Reinforcement Learning, and Ensemble Methods. This paper summarizes the benefits and drawbacks of their use of smart grid cyber security, and further makes a qualitative comparison of the three methods from multiple performance indicators.

Multi-objective maintenance strategy for corroded pipelines considering the correlation of different failure modes

Corrosion is the leading cause of leakage and burst failures in long-distances gas pipelines. Periodic maintenance is often used to extend the service life of the pipeline. The presence of multiple performance indicators with competing objectives needs to be taken into account for pipeline maintenance strategy. To address these challenges, this paper proposes an approach based on multi-objective maintenance optimization. The copula function is applied to characterize the correlation of failure modes (leakage and burst). With the aims to optimize availability and cost rate, a multi-objective maintenance optimization model is established, and it is solved with the non-dominated sorting genetic algorithm II (NSGA II). The results show that the Pareto frontier of multi-objective optimization can provide more feasible maintenance schemes. Moreover, the effects of major variables on the availability and cost rate are investigated. It is found that compared with other parameters, downtime cost has the greater impact on performance indicators. The proposed method can provide theoretical reference for pipeline integrity management, thereby ensuring that the pipeline is in a safe state during operation.

Introduction of ultra-hypofractionation in breast cancer: Implications for costs and resource use

PurposeA shift towards (ultra-)hypofractionated breast irradiation can have important implications for the practice of contemporary radiation oncology. This paper presents a systematic analysis of the impact of different fractionation schedules on multiple key performance indicators, namely resource use, costs, work times, throughput and waiting times. Materials and MethodsTime-driven activity-based costing (TD-ABC) is applied to calculate the costs and resources consumed where the perspective of the radiotherapy department in adopted. Three fractionation regimens are considered: ultra-hypofractionation (5 x 5.2 Gy, UHF), moderate hypofractionation (15 x 2.67 Gy, HF) and conventional fractionation (25 x 2 Gy, CF). Subsequently, a discrete event simulation (DES) model of the radiotherapy care pathway is developed and scenarios are compared in which the following factors are varied: distribution of fractionation regimens, patient volume and operating hours. ResultsThe application of (U)HF can permit radiotherapy departments to reduce the use of scarce resources, realise work time and cost savings, increase throughput and reduce waiting times. The financial advantages of (U)HF are, however, reduced in cases of excess capacity and cost savings may therefore be limited in the short-term. Moreover, although an extension of operating hours has favourable effects on throughput and waiting times, it may also reduce cost differences between fractionation schedules by increasing the capacity of resources. ConclusionBy providing an in-depth analysis of the consequences associated with a shift towards (U)HF in breast cancer, the present study demonstrates how a DES model based on TD-ABC costing can assist radiotherapy professionals in making data-driven decisions.

A Multi-Fidelity Integration Method for Reliability Analysis of Industrial Robots

Abstract A multi-fidelity integration method is proposed to analyze the reliability of multiple performance indicators (MPI) for industrial robots. In order to high-fidelity mapping the performance of industrial robots, a unified multi-domain model (UMDM) is first established. The contribution-degree analysis is then used to classify the input random variables into interacting and non-interacting ones. Thus, the high-dimensional integration of reliability analysis is separated into a low-dimensional integration and multiple one-dimensional integrations in an additive form. Here, the low-dimensional integration consisting of the interacting variables is calculated using the high-precision mixed-degree cubature formula (MDCF), and the computational results are treated as high-fidelity data. The one-dimensional integration consisting of non-interacting variables is then computed by the highly efficient five-point Gaussian Hermite quadrature (FGHQ), and the computational results are named low-fidelity data. A multi-fidelity integration method is constructed by fusing the high-fidelity data and the low-fidelity data to obtain the statistical moments of the MPI. Subsequently, the probability density function and the failure probability of the MPI are estimated using the saddlepoint approximation method. Finally, some representative methods are performed to verify the superiority of the proposed method.

Dynamic Resource Allocation in Queue-Constrained and Delay-Sensitive Vehicular Networks

Since performing reliable and real-time data transmission is challenging, this paper focuses on the dynamic resource allocation framework with queue and delay constraints in vehicular relay networks, where vehicle users and roadside users are in cooperative communications. To reduce data fluctuation and tackle channel uncertainty, the Lyapunov optimization theory is integrated into the long-term dynamic resource allocation framework, which addresses multiple performance indicators including queue backlog, communication delay, and user utility. In particular, the hierarchical competitive relationship between vehicle users and roadside users is captured, and the Stackelberg game structure is adopted to maximize the utilities of all users. Furthermore, the lower-game subproblem is transformed by the Lyapunov optimization method and is solved by the Lagrangian function method. As the best response function, the lower-game solution is used to formulate the upper-game strategy. Since the existing of game equilibrium (GE) solution, a Stackelberg game-enabled robust resource allocation algorithm can be proposed. Finally, various numerical simulations are performed to show that the proposed algorithm has better robustness and higher throughout than the benchmarks.

Remaining useful life prediction for complex systems with multiple indicators of stochastic correlation considering random shocks

The remaining useful life (RUL) prediction in complex systems is subject to complex working conditions, such as overloads, random vibration and shocks, gradual corrosion and wearing, that cause the complex and variable performance degradation of the mechanical system. Accurate prediction of RUL requires the extraction of multiple performance indicators strongly correlated to system degradation. Accordingly, we propose an RUL prediction framework for a multi-indicator system that considers the effects of random shocks and stochastic dependence. First, considering the influence of random shocks and measurement errors, a degradation state-space model is established using the stochastic correlation of multiple performance indicators. Next, the hidden state and unknown parameters of the state-space model are jointly estimated using the expectation–maximization algorithm in conjunction with the strong tracking filter algorithm based on online monitoring data. Subsequently, the probability density function expression of the RUL is derived for a multi-indicator system with three failure modes. Finally, the accuracy of the RUL prediction model is verified for a system affected by random shocks via numerical experiments. Compared with other prediction methods, the effectiveness and adaptability of the proposed method are verified using the C-MAPSS dataset and a high-temperature furnace case.

A multidimensional framework for asphalt pavement evaluation based on multilayer network representation learning: A case study in RIOHTrack

Pavement evaluation using multiple performance indicators has been a critical challenge in the field due to limitations in relying on pavement engineers to simultaneously assess several performance attributes, which usually leads to subjectivity and variability. In this study, a data-driven multidimensional framework is proposed to compact this issue by utilizing multilayer network representation learning. The key to this framework is to capture not only the performance conditions per se, but also the nonlinear interactions among these attributes. This provides an in-depth higher-order property of pavements’ service condition. Specifically, pavement performance attributes are modeled into multilayer network with each layer representing an aspect of pavement condition. Subsequently, this multilayer pavement condition network is mapped into low-dimensional space through the network representation learning for systematic evaluation. Finally, unsupervised cluster analysis derives groups of pavements which share similar overall condition for future decision-making process. The proposed method is validated with a case study in the Research Institute of Highway Ministry track (RIOHTrack) and experimental results demonstrate the effectiveness in categorizing pavement condition based on multi-attributes.

Performance evaluation of deep learning algorithms for heat loss damage classification in buildings from UAV-borne infrared images

Deep learning (DL) algorithms are gaining popularity to automate anomaly detection in a variety of civil and mechanical engineering domains. Nevertheless, very few attempts have been made to use infrared (IR) images with DL algorithms to detect heat loss in a building and expedite the otherwise cumbersome energy audit process. However, no clear guidelines exist for the selection of appropriate algorithms or hyperparameters (e.g., number of epochs, learning rate, number of images) to implement DL with IR images. To address the research gap, a comparative sensitivity analysis of four well-known DL algorithms for the classification of heat losses using only IR images was performed for the first time. Four DL algorithms, namely, general convolutional neural network (CNN), VGG16, transfer learning (TL) VGG16, and TL inceptionV3 were selected and used for automating the detection of heat loss from IR images. The sensitivity of the selected CNN algorithms was analyzed based on i) the size of the IR images dataset, ii) the number of epochs, and iii) the effect of an unbalanced dataset. The goal was to provide a framework that may aid in the selection of suitable CNN architecture as well as parameters related to data size and an optimal number of training iterations for reliable heat loss identification using IR images. This article has provided a framework for automated heat loss detection that can be applied to any IR dataset without the need to pre-process the IR images. The accuracy of the selected DL algorithms is validated with IR images of laboratory setups and real-world buildings characterized by different types of defects. The comparative sensitivity analysis was conducted based on multiple performance indicators such as precision, recall, and F-1 score which resulted in optimal values of hyperparameters and the most suitable DL algorithm. The automated identification of heat loss yielded performance indicator values of more than 90%.

Determining safety aspects of Differential Speed Limit on Indian Expressway

Optimization of operating speed limit on any class of road is the biggest task faced by the transportation experts. A variety of speed control or management techniques are available all around the globe to minimize the frequency of over-speeding or for increasing the acceptability of the posted speed limit. In this study, a comparison is carried out between Uniform Speed Limit (USL) and Differential speed Limit (DSL) on an Expressway in India, using microsimulation package PTV VISSIM. In USL, a single speed limit is implemented for all class of vehicles whereas DSL strategy have two distinct speed limit for two different class of vehicles i.e., passenger carrying vehicles and goods carrying vehicles. Multiple Key Performance Indicators (KPIs) were considered. Results suggested that DSL seems to be the better strategy on the roads accommodating high speeds like Expressways under heterogenous conditions. Four out of five KPIs were showing the results in favor of DSL strategy in comparison to USL strategy. KPIs like speed variance, Road Safety Index (RSI), no. of vehicles having Deceleration-Rate-to-Avoid-Collision (DRAC) > 3.35 m/sec2 and no. of lane changes were showcasing better results in case of DSL strategy in comparison to USL strategy.

Performance analysis of rectangular and double-D transmitters with various receivers for electric vehicle static charging

The inductive charging stations of the future should be able to charge any car, regardless of the type of coil configuration attached to it. In this study the performance of different configurations of coils is investigated. Where the principle of interoperability is studied between the most common types of coils in the IPT system (rectangular and double-D) recommended by Society of Automotive Engineering (SAE) J2954. This interoperability analysis is done at a power level of 11.1 kVA and Z3-class ground clearance using the two types of coils on the side of the car, while double D (DD) is used as a general transmitter in one case, and in the other case, a rectangle (R) is used as a general transmitter. Four 3D finite-element models (FEMs), along with compensation circuits, are built to perform the interoperability tests. The circuit models are utilized to determine the frequency of operation and the parameters of compensation circuit that realize the rated power at highest efficiency. The concept of interoperability has been verified through multiple performance indicators, namely coupling coefficient, transmitted power, and transmission efficiency. These indicators were calculated at ideal alignment and the different cases of misalignment, whether lateral or angular. The findings prove that the two different types of receivers (R and DD) can work efficiently and smoothly with the general transmitter, whether DD pad or rectangular pad are used. At the same time, the nominal power can be transmitted at high efficiency, above the limits permitted by SAE j2954 (>85 % at alignment conditions and >80 % at misalignment conditions).

Semantic Communication-Based Dynamic Resource Allocation in D2D Vehicular Networks

The semantic communication mechanism enables wireless devices in vehicular networks to communicate more effectively with the semantic meaning. However, in high-dynamic vehicular networks, the transmission of semantic information faces challenges in terms of reliability and stability. To address these challenges, a long-term robust resource allocation scheme is proposed under the Device-to-Device (D2D) vehicular (D2D-V) networks, where multiple performance indicators (user satisfaction, queue stability, and communication delay) are considered. Due to the sophisticated probabilistic form with consideration of channel fluctuations, the Bernstein approximation is introduced to acquire the deterministic constraint more efficiently. The robust resource allocation problem is proposed and separated into two independent subproblems by the Lyapunov optimization method, which includes semantic access control in the application layer and power control in the physical layer. After that, the successive convex approximation method and Karush-Kuhn-Tucher conditions are adopted to solve the subproblems, thereby proposing a robust resource allocation algorithm. The simulations reveal the trade-off relationship between user satisfaction, queue stability, and communication delay, which is on the premise of meeting the user SINR requirement. Moreover, the simulations also prove the necessity of considering channel uncertainty in high-speed mobile vehicular communication scenarios.