Heavy System Load Research Articles

135 Acceptance and Barriers of AI-Based Decision Support Systems in Emergency Departments: A Quantitative and Qualitative Evaluation

Computerized Decision Support Systems (CDSSs) are developed to assist clinicians in their decision-making processes with the ultimate goal of improving patients’ outcomes. Although many studies have discussed the acceptance and barriers of CDSSs, few have been investigated with theoretically grounded frameworks, especially AI-based CDSSs. We aimed to identify the acceptance and barriers of our newly developed CDSS by using the Unified Theory of Acceptance and Use of Technology (UTAUT) and the Consolidated Framework for Implementation Research (CFIR) frameworks. This mixed-methods study was performed from March to April 2021. Transitional year residents (n=6), emergency medicine residents (n=5), and emergency physicians (n=3) from two community, tertiary care hospitals in Japan were included. We first developed a real-time CDSS for predicting aortic dissection based on numeric and text information from medical charts (eg, chief complaints, medical history, vital signs) with natural language processing. To address anchoring bias, the system displays an alert with an explanation on the screen when the probability of aortic dissection changes sharply or exceeds the prespecified threshold. This system was deployed on the Internet, and the participants used the system with clinical vignettes based on actual cases. Participants were then involved in a mixed-methods evaluation consisting of a UTAUT-based questionnaire with a 5-point Likert scale (quantitative) and a CFIR-based semi-structured interview (qualitative). Interviews were sampled, transcribed, and analyzed using the MaxQDA software. A deductive approach to qualitative content analysis was adopted incorporating four of the CFIR’s domains. The qualitative data were interpreted through coding and a summary matrix of the domains. All 14 participants completed the questionnaires and interviews. Quantitative analysis revealed generally positive responses for user acceptance. Effort expectancy (median [IQR], 4 [3-4]; 5 indicates most positive) was especially high which led to high behavior intention (median [IQR], 3 [3-4]). Qualitative analysis suggested that the design quality was high and was perceived to be neither interruptive nor effortful. In addition, it implied that the system was reliable enough. Despite its reliability and ease of use, alerts by the CDSS tended to be ignored for patients with typical presentations, for whom participants felt more confident about the diagnostic algorithms. Moreover, while real-time updates of alerts were perceived useful, it led to concerns that the heavy system load may lead to errors in the system. Using theoretically grounded frameworks enabled us to thoroughly understand the acceptance and barriers of implementation of CDSS. Although our real-time CDSS was qualitatively shown to be less effective on typical cases with concerns of system failure, the mixed-methods evaluation demonstrated that the CDSS was intuitive and reliable and thus has good user acceptance.

Collaborative Learning Based Straggler Prevention in Large-Scale Distributed Computing Framework

Modern big data applications tend to prefer a cluster computing approach as they are linked to the distributed computing framework that serves users jobs as per demand. It performs rapid processing of tasks by subdividing them into tasks that execute in parallel. Because of the complex environment, hardware and software issues, tasks might run slowly leading to delayed job completion, and such phenomena are also known as stragglers. The performance improvement of distributed computing framework is a bottleneck by straggling nodes due to various factors like shared resources, heavy system load, or hardware issues leading to the prolonged job execution time. Many state-of-the-art approaches use independent models per node and workload. With increased nodes and workloads, the number of models would increase, and even with large numbers of nodes. Not every node would be able to capture the stragglers as there might not be sufficient training data available of straggler patterns, yielding suboptimal straggler prediction. To alleviate such problems, we propose a novel collaborative learning-based approach for straggler prediction, the alternate direction method of multipliers (ADMM), which is resource-efficient and learns how to efficiently deal with mitigating stragglers without moving data to a centralized location. The proposed framework shares information among the various models, allowing us to use larger training data and bring training time down by avoiding data transfer. We rigorously evaluate the proposed method on various datasets with high accuracy results.

Performance-aware load shedding for monitoring events in container based environments

Runtime monitoring tools have become fundamental to assess the correct operation of complex systems and applications. Unfortunately, the more precise is the monitoring (sampling rate, information granularity, and so on), the higher is the overhead introduced in the system itself. In this paper, we propose a new load shedding framework that enables runtime adaptation of monitoring agents under heavy system load, exploiting an heuristic Load Manager to control the agent status and a runtime support for domain-specific policies. We implemented the proposed methodology on Sysdig, with an average control error improvement of 3.51x (12.25x at most), w.r.t. previous solutions.

Raise Voltage Stability Limit of a Power System using Reactive Power Compensation Technique

In recent years, voltage stability problems have been increasing since power systems operate close to stability limits. The voltage stability problem of a power system is associated with a rapid voltage drop due to heavy system load and it occurs because of inadequate reactive power support at some critical bus. One of the serious consequences of the voltage stability is a system blackout, and this has received more attention in recent years. Accurate determination of stability limit and amount of reactive power injection to stabilize is important.This paper proposes to determine voltage stability margin of a critical bus and also provide amount of reactive power injection to the bus particularly during overload, a simple two bus equivalent model of the power system is used to determine the maximum apparent power for different power factors. Any required apparent power can directly obtained by correcting the reactive power at critical bus. Experimental results support our theoretical findings.

Raise Voltage Stability Limit of a Power System using Reactive Power Compensation Technique

In recent years, voltage stability problems have been increasing since power systems operate close to stability limits. The voltage stability problem of a power system is associated with a rapid voltage drop due to heavy system load and it occurs because of inadequate reactive power support at some critical bus. One of the serious consequences of the voltage stability is a system blackout, and this has received more attention in recent years. Accurate determination of stability limit and amount of reactive power injection to stabilize is important.This paper proposes to determine voltage stability margin of a critical bus and also provide amount of reactive power injection to the bus particularly during overload, a simple two bus equivalent model of the power system is used to determine the maximum apparent power for different power factors. Any required apparent power can directly obtained by correcting the reactive power at critical bus. Experimental results support our theoretical findings.

The +1 Method: Model-Free Adaptive Repositioning Policies for Robotic Multi-Agent Systems

Robotic multi-agent systems can efficiently handle spatially distributed tasks in dynamic environments. Problem instances of particular interest, and generality are the dynamic traveling repairman problem, and the dynamic vehicle routing problem. Operational policies for robotic fleets solving these two problems take decisions in an online setting with continuously arriving demands to optimize service level, and efficiency, and can be classified along several lines. First, some require a model of the demand, e.g., based on historical information, while others work model-free. Second, they are designed for different operating conditions from light to heavy system load. Third, they work in a time-invariant or time-varying setting. We present a novel class of model-free operational policies for time-varying demands, which have performance independent of the load factor, a combination of properties not achieved by other operational policies in the literature. The underlying principle of the introduced policies is to send available robots to recent service request locations. In simple terms, they rely on sending more than one robot for every service request arriving to the system. This leads to an advantage in scenarios where demand is non-uniformly distributed, and correlated in space, and time. We provide performance guarantees for both the time-invariant, and the time-varying cases as well as for correlated demand. We verify our theoretical results numerically. Finally, we apply our operational policy to the problem of mobility-on-demand fleet operation, and demonstrate that it outperforms model-based, and complex algorithms across all load ranges, despite its simplicity.

A multi-objective optimization for HAWT blades design by considering structural strength

The challenge of wind turbine blade design is to balance the conflict between high capacity and heavy system loads introduced by the large scale rotor. To solve this problem, we present a multi-objective optimization method to maximize the Annual energy production (AEP) and minimize the blade mass. The well-known Blade element momentum (BEM) theory is employed to predict the aerodynamic performance and AEP of the blade. The blade is simplified as a thin Bernoulli beam. The cross section is modelled as a combination of composite layer, shear webs and spar caps typically. The strain of every cross section has been considered as a constraint to minimize the spar cap thickness for minimizing the blade mass. An improved genetic algorithm (NSGA-II) is applied to obtain the Pareto front set. Several solutions of the Pareto set are selected to compare with the reference blade (NREL 5MW blade). Performance of the rotors on design condition is simulated by STAR-CCM+ to verify the results of BEM theory. Optimal results show that the present blade, which is fully superior to the reference blade, can be selected from the Pareto set. The optimization design method can provide a superior blade with an increase by 2.48% of AEP and a reduction by 5.52% of the blade mass. It indicates the present optimization method is effective. Results of numerical simulations show that the spanwise flow would be increased obviously in tip region of the reference blade. The reason is that chord length variation in blade tip affects the flow and causes minor stall. The abrupt change of chord distribution in blade tip should be avoided to reduce the spanwise flow in initial blade design.

Performance of a novel evolutionary genetic-based multi-user detector for multi-carrier CDMA communication systems

In this paper, we investigate the performance of a modified evolutionary multi-user detector (MUD) in multi-carrier direct-sequence code-division multiple-access (MC-CDMA) communication systems over frequency-selective fading channels. The genetic algorithm (GA), which tries to reduce the complexity in optimal detection due to greedy exhausted searching, performs using the adequate sub-optimal multi-user detector for MC-CDMA systems. In this paper, we propose a forcing mutation scheme for the genetic algorithm (GA) to perform a forcing mutation evolutionary multi-user detector (FME-MUD) and to largely reduce computational complexity. Simulation results show that with six generations, the proposed FME-MUD scheme outperforms the conventional GA-MUD, especially at heavy system loads.

Voltage Stability Assessment in Radial Distribution Power System

The voltage stability problem of distribution networks is associated with a rapid voltage drop because of heavy system load. To operate the distribution system under such critical conditions the integration of distributed resource improves the reliability of supplying power by improving voltage stability and reducing the power losses. This paper presents voltage stability analysis of radial distribution networks in the presence of distributed generation. The analysis is accomplished using different methods which can be evaluated at each node of the distribution system. In this paper study the all types of methods which is used to improve the Voltage Stability Assessment in radial distribution power system.

An Automatic Decision-Making Mechanism for Virtual Machine Live Migration in Private Clouds

Due to the increasing number of computer hosts deployed in an enterprise, automatic management of electronic applications is inevitable. To provide diverse services, there will be increases in procurement, maintenance, and electricity costs. Virtualization technology is getting popular in cloud computing environment, which enables the efficient use of computing resources and reduces the operating cost. In this paper, we present an automatic mechanism to consolidate virtual servers and shut down the idle physical machines during the off-peak hours, while activating more machines at peak times. Through the monitoring of system resources, heavy system loads can be evenly distributed over physical machines to achieve load balancing. By integrating the feature of load balancing with virtual machine live migration, we successfully develop an automatic private cloud management system. Experimental results demonstrate that, during the off-peak hours, we can save power consumption of about 69 W by consolidating the idle virtual servers. And the load balancing implementation has shown that two machines with 80% and 40% CPU loads can be uniformly balanced to 60% each. And, through the use of preallocated virtual machine images, the proposed mechanism can be easily applied to a large amount of physical machines.

Voltage stability analysis of radial distribution networks using catastrophe theory

The voltage stability problem of distribution networks is associated with a rapid voltage drop because of heavy system load. The decline of voltage stability level will restrict the increase of load served by distribution companies. One of the serious consequences of the voltage stability problem is a system blackout, and this problem has received much attention of distribution companies. In this study, a new voltage stability index (VSI) is developed for identifying the most sensitive bus to the voltage collapse in radial distribution network. The analysis is based on the catastrophe theory to find a new VSI and the voltage stability boundary of a distribution system. Two tested 15- and 30-bus systems analysed to give an illustration of the feasibility of the proposed method.

A new window-based job scheduling scheme for 2D mesh multicomputers

Allocating submeshes to jobs in mesh-connected multicomputers in a FCFS fashion can lead to poor system performance (e.g., long job waiting delays) because the job at the head of the waiting queue can prevent the allocation of free submeshes to other waiting jobs with smaller submesh requirements. However, serving jobs aggressively out-of-order can lead to excessive waiting delays for jobs with large allocation requests. In this paper, we propose a scheduling scheme that uses a window of consecutive jobs from which it selects jobs for allocation and execution. This window starts with the current oldest waiting job and corresponds to the lookahead of the scheduler. The performance of the proposed window-based scheme has been compared to that of FCFS and other previous job scheduling schemes. Extensive simulation results based on synthetic workloads and real workload traces indicate that the new scheduling strategy exhibits good performance when the scheduling window size is large. In particular, it is substantially superior to FCFS in terms of system utilization, average job turnaround times, and maximum waiting delays under medium to heavy system loads. Also, it is superior to aggressive out-of-order scheduling in terms of maximum job waiting delays. Window-based job scheduling can improve both overall system performance and fairness (i.e., maximum job waiting delays) by adopting large lookahead job scheduling windows.

A Load Sharing Algorithm Including An Improved Response Time using Evolutionary Information in Distributed Systems

A load sharing algorithm is one of the important factors in computer system. In sender-initiated load sharing algorithms, when a distributed system becomes to heavy system load, it is difficult to find a suitable receiver because most processors have additional tasks to send. The sender continues to send unnecessary request messages for load transfer until a receiver is found while the system load is heavy. Because of these unnecessary request messages it results in inefficient communications, low cpu utilization, and low system throughput. To solve these problems, we propose a self-adjusting evolutionary algorithm for improved sender-initiated load sharing in distributed systems. This algorithm decreases response time and increases acceptance rate. Compared with the conventional sender-initiated load sharing algorithms, we show that the proposed algorithm performs better.

An adaptive multiuser detector for DS‐CDMA systems in multipath fading channels

AbstractIn this paper, we investigate the performance of an adaptive multistage detection scheme for direct‐sequence code‐division multiple‐access (DS‐CDMA) systems. The first stage consists of an adaptive multiuser detector which is based on the linear constrained minimum variance (LCMV) criterion. The interference cancellation (IC) occurs in the second stage. The performance of the iterative receiver over both flat and frequency‐selective fading channels is investigated and compared to the single‐user bound. In all cases, and under heavy system loads with near‐far problems, the iterative receiver is shown to offer substantial performance improvement and large gain in user‐capacity relative to the standard LCMV. In flat‐fading channels, our results show that the performance of the iterative detector is very close to the single‐user bound. For the frequency‐selective channel, this performance is noted to be in the order of 1 dB far from the single‐user bound. Copyright © 2007 John Wiley & Sons, Ltd.

다중 분할된 구조를 가지는 클러스터 검사점 저장 기법

검사점 저장 기법을 사용하여 주기적으로 클러스터 노드들의 프로세스 수행 정보를 전역 저장 장치에 저장하는 분산 클러스터 시스템에서 결함 허용 성능을 유지하는 데 드는 비용을 줄이고 전체 프로세스의 수행 성능을 증가시키기 위해서는 검사점 정보를 저장할 때에 네트워크로 전달되는 부하를 각 노드에 최대한 적절하게 분산하여 데이터 저장 시간을 줄임으로써 검사점 정보를 저장하는 동안 전체 클러스터 시스템의 프로세스가 지연되는 시간을 줄이도록 하여야 한다. 이를 위하여 분산 RAID 기반의 단일 입출력 공간을 사용하는. 클러스터 시스템에서는 여러가지 검사점 저장 기법을 사용하며, 검사점 정보의 저장 기법에 따라서 저장 성능과 결함 회복 성능이 달라진다. 본 연구에서는 분할된 검사점 저장 기법을 개선하여 검사점 데이터를 분산 RAID 기반의 단일 입출력 공간에 저장할 때에 그룹별로 분할되는 분할 그룹 크기를 검사점 정보가 저장될 때의 네트워크의 트래픽에 따라서 동적으로 결정하여 네트워크를 통한 분산 RAID에 저장함으로써 네트워크 병목현상을 최소화하는 다중 분할된 검사점 저장 구조를 제안하였다. 제안된 구조의 성능을 분석하기 위하여 최대 512개의 가상 노드로 구성된 클러스터 시스템을 대상으로 하여 MPI 와 Linpack HPC 벤치마크를 통한 성능 평가를 수행하였으며, 성능 평가 결과는 검사점 정보의 크기와 클러스터의 크기가 증가할수록 제안된 기법이 검사점 정보의 저장과 결함 회복 능력에 대하여 기존의 검사점 저장 기법에 비하여 우수한 성능을 보인다. The checkpointing schemes should reduce the process delay through managing the checkpoints of each node to fit the network load to enhance the performance of the process running on the cluster system that write the checkpoints into its global stable storage. For this reason, a cluster system with single IO space on a distributed RAID chooses a suitable checkpointng scheme to get the maximum IO performance and the best rollback recovery efficiency. In this paper, we improved the striped checkpointing scheme with dynamic stripe group size by adapting to the network bandwidth variation at the point of checkpointing. To analyze the performance of the multi striped checkpointing scheme, we applied Linpack HPC benchmark with MPI on our own cluster system with maximum 512 virtual nodes. The benchmark results showed that the multistriped checkpointing scheme has better performance than the striped checkpointing scheme on the checkpoint writing efficiency and rollback recovery at heavy system load.

A hybrid approach to adaptive load sharing and its performance

Abstract The average response time of tasks in a distributed system depends on the strategy by which workload is shared among the nodes of the system. A common approach to load sharing is to resort to some distributed algorithm that arranges for task transfer between nodes based on information on the system's state. In this paper, we depict a hybrid approach to adaptive load sharing which outperforms existing algorithms, and is especially effective in response to peaks of workload, under both heavy and light system load conditions. The strategy we propose is novel in that it relies on a fully distributed algorithm when the system is heavily loaded, but resorts to a centrally coordinated one when parts of the system become idle. The transition from one algorithm to the other is performed automatically, and the simplicity of the algorithms proposed makes it possible to use a centralized component without incurring in scalability problems and presenting instabilities. Both algorithms are very lightweight and do not need any tuning of parameters. Simulations show that the hybrid approach performs well under all load conditions and task generation patterns, it is weakly sensitive to processing overhead and communication delays, and scales well (to hundred of nodes) despite the use of a centralized component.

A fast method for determining the voltage stability limit of a power system

Abstract The voltage stability problem of a power system is associated with a rapid voltage drop due to heavy system load, and it occurs because of inadequate reactive power support at some critical buses. One of the serious consequences of the voltage stability problem is a system blackout, and this problem has received much attention in recent years. This paper proposes a fast method for finding the maximum load, especially the reactive power demand, at a particular load bus before reaching the voltage stability limit. The method uses the base-case system information to find special two-bus equivalent of the system for analyzing the voltage stability problem. The method was tested on the IEEE 14-, 30-, and 118-bus systems and the results obtained were compared with those found by some other methods.