Adaptive Workload Management Research Articles

Coordination breakdowns in nuclear power plant control rooms: cause identification and behaviour-sequence analysis

This study aims to identify the causes of coordination breakdowns among control crews and to understand their coordination-behaviour patterns during emergencies in nuclear power plants (NPPs). On the basis of in-depth interviews with 18 control-crew operators, we identified 25 causes of coordination breakdown related to work processes, personnel, and situation and organisation. In addition, we observed 12 control-crew training sessions that dealt with emergencies and conducted lag-sequential analysis. The levels of coordination effectiveness were evaluated using the proportion of coordination breakdowns and the anticipation ratio. We found that higher-performing teams exhibited more non-random coordination behavioural patterns than did lower-performing teams. Coordination-behaviour patterns specific to the higher-performing teams included adaptive workload management (from senior operators) and proactive seeking performance monitoring (from junior operators). The findings of the study enrich our understanding of the critical factors and processes that influence coordination effectiveness of NPP control crews. Practitioner summary: Causes of coordination breakdowns among control crews of NPPs were identified based on in-depth interviews with control-crew operators, and behavioural-pattern analysis of control crews in 12 training sessions were analysed to reveal the patterns that differentiate higher- and lower-performing teams. The findings of the study enrich our understanding of the critical factors and processes that influence the coordination effectiveness of NPP control crews. Abbreviations: NPP: nuclear power plant; RO: reactor operator; TO: turbine operator; CO: coordinator; SRO: senior reactor operator

Adaptive Workload Management in Mixed-Criticality Systems

Due to the efficient resource usage of integrating tasks with different criticality onto a shared platform, the integration with mixed-criticality tasks is becoming an increasingly important trend in the design of real-time systems. One challenge in such a mixed-criticality system is to maximize the service for low-critical tasks, while meeting the timing constraints of high-critical tasks. In this article, we investigate how to adaptively manage the low-critical workload during runtime to meet both goals, that is, providing the service for low-critical tasks as much as possible and guaranteeing the hard real-time requirements for high-critical tasks. Unlike previous methods, which enforce an offline bound towards the low-critical workload, runtime adaptation approaches are proposed in which the incoming workload of low-critical tasks is adaptively regulated by considering the actual demand of high-critical tasks. This actual demand of the high-critical tasks, in turn, is adaptively updated using their historical arrival information. Based on this adaptation scheme, two scheduling policies—the priority-adjustment policy and the workload-shaping policy—are proposed to do the workload management. In order to reduce online management overhead, a lightweight scheme with O ( n · log ( n )) complexity is developed. Extensive simulation results are presented to demonstrate the effectiveness of our proposed workload management approaches.

Mathematical Modeling of the Human Cognitive System in Two Serial Processing Stages With Its Applications in Adaptive Workload-Management Systems

With the increasing usage of in-vehicle systems, drivers have to frequently perceive and respond to messages from these in-vehicle systems. In addition, previous studies have found that the interval between the messages (arrival rate) presented to a driver becomes one of the factors affecting driver workload. To reduce driver workload, researchers on adaptive workload-management systems have found that adding extra delay time into the interval of messages can significantly reduce driver workload. However, it is unknown whether this extra delay time added by an adaptive workload-management system will increase the performance time of drivers or not. To answer this important question, using closed-form mathematical equations, the current work quantifies human performance time (total task completion time and reaction time of each task) when there are two serial processing stages in the human cognitive system. The mathematical model developed in this work provides solutions of the optimal interval of messages that generate the lowest workload without deteriorating drivers' performance time to respond to multiple messages from in-vehicle systems. This is one of a few closed-form deterministic mathematical models with analytic solutions that can predict average reaction time when there are two multiple serial stages in the cognitive system in dual tasks. With relatively simple equations, the mathematical model can still capture the major patterns of simulation results with stochastic properties and human behavioral experimental results. The mathematical equations developed in this study can be used in the design of adaptive workload-management systems and other driver assistance systems.

Development of an Adaptive Workload Management System Using the Queueing Network-Model Human Processor (QN-MHP)

The risk of vehicle collisions significantly increases when drivers are overloaded with information from in-vehicle systems. One of the solutions to this problem is developing adaptive workload management systems (AWMSs) to dynamically control the rate of messages from these in-vehicle systems. However, existing AWMSs do not use a model of the driver cognitive system to estimate workload and only suppress or redirect in-vehicle system messages, without changing their rate based on driver workload. In this paper, we propose a prototype of a new queueing network-model human processor AWMS (QN-MHP AWMS), which includes a queueing network model of driver workload that estimates the driver workload in several driving situations and a message controller that determines the optimal delay times between messages and dynamically controls the rate of messages presented to drivers. Given the task information of a secondary task, the QN-MHP AWMS adapted the rate of messages to the driving conditions (i.e., speeds and curvatures) and driver characteristics (i.e., age). A corresponding experimental study was conducted to validate the potential effectiveness of this system in reducing driver workload and improving driver performance. Further development of the QN-MHP AWMS, including its use in in-vehicle system design and possible implementation in vehicles, is discussed.

Development of an Adaptive Workload Management System using the Queueing Network-Model Human Processor

Drivers overloaded with information from in-vehicle systems significantly increase the chance of vehicle collisions. Developing adaptive workload management systems (AWMS) to dynamically control the rate of messages from these in-vehicle systems is one of the solutions to this problem. However, existing AWMS do not use driver models to estimate workload, and only suppress or redirect messages without changing the rate of messages from the in-vehicle systems. In this work, we propose a prototype of a new adaptive workload management system, the Queuing Network-Model Human Processor (QN-MHP) AWMS, which includes a model of driver workload based on the queueing network theory of human performance that estimates driver workload in different driving situations and a message controller that dynamically controls the rate of messages presented to drivers. Corresponding experimental studies were conducted to validate the potential effectiveness of this system in reducing driver workload and improving driver performance.

Adaptive Workload Management through Elastic Scheduling

In real-time computing systems, timing constraints imposed on application tasks are typically guaranteed off line using schedulability tests based on fixed parameters and worst-case execution times. However, a precise estimation of tasks’ computation times is very hard to achieve, due to the non-deterministic behavior of several low-level processor mechanisms, such as caching, prefetching, and DMA data transfer. The disadvantage of relying the guarantee test on a priori estimates is that an underestimation of computation times may jeopardize the correct behavior of the system, whereas an overestimation will certainly waste system resources and causes a performance degradation. In this paper, we propose a new methodology for automatically adapting the rates of a periodic task set without forcing the programmer to provide a priori estimates of tasks’ computation times. Actual executions are monitored by a runtime mechanism and used as feedback signals for predicting the actual load and achieving rate adaptation. Load balancing is performed using an elastic task model, according to which tasks utilizations are treated as springs with given elastic coefficients.