Commercial servers, such as database or application servers, often attempt to improve performance via multithreading. Improper multi-threading architectures can incur contention, limiting performance improvements. Contention occurs primarily at two levels: (1) blocking on locks shared between threads at the software level and (2) contending for physical resources (such as the cpu or disk) at the hardware level. Given a set of hardware resources and an application design, there is an optimal number of threads that maximizes performance. This paper describes a novel technique we developed to select the optimal number of threads of a target-tracking application using a simulationbased Colored Petri Nets (CPNs) model. This paper makes two contributions to the performance analysis of multi-threaded applications. First, the paper presents an approach for calibrating a simulation model using training set data to reflect actual performance parameters accurately. Second, the model predictions are validated empirically against the actual application performance and the predicted data is used to compute the optimal configuration of threads in an application to achieve the desired performance. Our results show that predicting performance of application thread characteristics is possible and can be used to optimize performance.
Standardizing the care of patients with complex problems in hospital settings is a challenge for physicians, nurses and other medical professionals. In acute care settings such as intensive care units, the inherent problems of stabilizing and improving vital patient parameters is complicated by the division of responsibilities among different individuals and teams. The use of evidence-based guidelines for managing complex clinical problems has become the standard of practice. Computerized support for implementing such guidelines has tremendous potential. The use of model-based techniques for specifying and implementing guidelines as coordinated asynchronous processes is a promising new methodology for providing advanced clinical decision support. Combined with visual dashboards, which show the status of the implemented guidelines, a new approach to computer-supported care is possible. These techniques are being applied to the management of sepsis in acute care settings at Vanderbilt Medical Center.
In a recent paper we have shown how wave variables can be used to interconnect passive plants with passive controllers such that the system remains l2-stable in spite of time-varying delays and data dropouts. The present paper further enhances these results by providing a detailed model that captures time-varying delays, data dropouts and network capacity for wireless ring token networks. It also provides a new theorem showing how an asynchronous controller can be implemented, which maintains an l2-stable system. Simulations show that the asynchronous control of a passive motor reduces the overall distortion when compared with a synchronous controller which relies on lossy data reduction techniques. These two distinct results pave the way to study high-performance rate-adaptive control schemes that minimize their control rate in order to match the network capacity.
This paper presents a formal method to design a digital inertial control system for quad-rotor aircraft. In particular, it formalizes how to use approximate passive models in order to justify the initial design of passive controllers. Fundamental limits are discussed with this approach – in particular, how it relates to the control of systems consisting of cascades of three or more integrators in which input actuator saturation is present. Ultimately, two linear proportional derivative (PD) passive controllers are proposed to be combined with a nonlinear saturation element. It is also shown that yaw control can be performed independently of the inertial controller, providing a great deal of maneuverability for quad-rotor aircraft. A corollary, based on the sector stability theorem provided by Zames and later generalized for the multiple-input-output case by Willems, provides the allowable range of k for the linear negative feedback controller kI in which the dynamic system H1 : x1 -> y1 is inside the sector [a1, b1], in which −1 < a1, 0 < b1 <= 1, and b1 > a1. This corollary provides a formal method to verify stability, both in simulation and in operation for a given family of inertial set-points given to the quadrotor inertial controller. The controller is shown to perform exceptionally well when simulated with a detailed model of the STARMAC, which includes blade flapping dynamics.
Actuator constraints such as saturation can impose severe constraints on networked control systems. For instance delays in wireless control systems of unstable plants combined with actuator constraints may make it impossible to stabilize a system. In this paper the conditions are derived that show when actuator saturation, a common memoryless nonlinearity, in series with a passive system causes the loss of passivity. However, using a non-linear controller known as a inner-product recovery block, the overall passivity of the system is recovered. Furthermore, we note specific sector conditions in which strictly-input passivity and strictly-output passivity can be recovered. Finally, it is shown how the inner-product recovery block can be used to maintain an lm2-stable wireless control network.
Real-life cyber-physical systems, such as automotive vehicles, building automation systems, and groups of unmanned vehicles are monitored and controlled by networked control systems. The overall system dynamics emerges from the interaction among physical dynamics, computational dynamics, and communication networks. Network uncertainties such as time-varying delay and packet loss cause significant challenges. This paper proposes a passive control architecture for designing wireless networked control systems that are insensitive to network uncertainties. We describe the architecture for a system consisting of a robotic manipulator controlled by a digital controller over a wireless network and we show that the system is stable even in the presence of time-varying delays. We present simulation results that demonstrate the advantages of the architecture with respect to stability and performance and show that the system is insensitive to network uncertainties.
Objective: The goal of this research is to provide a framework to enable the model-based development, simulation, and deployment of clinical information system prototypes with mechanisms that enforce security and privacy policies. Methods: We developed the Model-Integrated Clinical Information System (MICIS), a software toolkit that is based on model-based design techniques and high-level modeling abstractions to represent complex clinical workflows in a service-oriented architecture paradigm. MICIS translates models into executable constructs, such as web service descriptions, business process execution language procedures, and deployment instructions. MICIS models are enriched with formal security and privacy specifications, which are enforced within the execution environment. Results: We successfully validated our design platform by modeling multiple clinical workflows and deploying them onto the execution platform. Conclusions: The model-based approach shows great promise for developing, simulating, and evolving clinical information systems with formal properties and policy restrictions.
The usability of model transformation languages depends on the level of abstractions one can work with in rules to perform complex operations on models. Recently, we have introduced a novel operator for our model transformation language GReAT that allows the concise specification of complex model (graph) rewriting operations that manipulate entire subgraphs. In this paper we show how the new operator can be used to implement non-trivial model manipulations with fewer and simpler rules, while maintaining efficiency. The examples were motivated by problems encountered in real-life model transformations.