Adaptive cache aware multiprocessor scheduling framework

Computer resource allocation represents a significant challenge particularly for multiprocessor systems, which consist of shared computing resources to be allocated among co-runner processes and threads. While an efficient resource allocation would result in a highly efficient and stable overall multiprocessor system and individual thread performance, ineffective poor resource allocation causes significant performance bottlenecks even for the system with high computing resources. This thesis proposes a cache aware adaptive closed loop scheduling framework as an efficient resource allocation strategy for the highly dynamic resource management problem, which requires instant estimation of highly uncertain and unpredictable resource patterns. Many different approaches to this highly dynamic resource allocation problem have been developed but neither the dynamic nature nor the time-varying and uncertain characteristics of the resource allocation problem is well considered. These approaches facilitate either static and dynamic optimization methods or advanced scheduling algorithms such as the Proportional Fair (PFair) scheduling algorithm. Some of these approaches, which consider the dynamic nature of multiprocessor systems, apply only a basic closed loop system; hence, they fail to take the time-varying and uncertainty of the system into account. Therefore, further research into the multiprocessor resource allocation is required. Our closed loop cache aware adaptive scheduling framework takes the resource availability and the resource usage patterns into account by measuring time-varying factors such as cache miss counts, stalls and instruction counts. More specifically, the cache usage pattern of the thread is identified using QR recursive least square algorithm (RLS) and cache miss count time series statistics. For the identified cache resource dynamics, our closed loop cache aware adaptive scheduling framework enforces instruction fairness for the threads. Fairness in the context of our research project is defined as a resource allocation equity, which reduces corunner thread dependence in a shared resource environment. In this way, instruction count degradation due to shared cache resource conflicts is overcome. In this respect, our closed loop cache aware adaptive scheduling framework contributes to the research field in two major and three minor aspects. The two major contributions lead to the cache aware scheduling system. The first major contribution is the development of the execution fairness algorithm, which degrades the co-runner cache impact on the thread performance. The second contribution is the development of relevant mathematical models, such as thread execution pattern and cache access pattern models, which in fact formulate the execution fairness algorithm in terms of mathematical quantities. Following the development of the cache aware scheduling system, our adaptive self-tuning control framework is constructed to add an adaptive closed loop aspect to the cache aware scheduling system. This control framework in fact consists of two main components: the parameter estimator, and the controller design module. The first minor contribution is the development of the parameter estimators; the QR Recursive Least Square(RLS) algorithm is applied into our closed loop cache aware adaptive scheduling framework to estimate highly uncertain and time-varying cache resource patterns of threads. The second minor contribution is the designing of a controller design module; the algebraic controller design algorithm, Pole Placement, is utilized to design the relevant controller, which is able to provide desired timevarying control action. The adaptive self-tuning control framework and cache aware scheduling system in fact constitute our final framework, closed loop cache aware adaptive scheduling framework. The third minor contribution is to validate this cache aware adaptive closed loop scheduling framework efficiency in overwhelming the co-runner cache dependency. The timeseries statistical counters are developed for M-Sim Multi-Core Simulator; and the theoretical findings and mathematical formulations are applied as MATLAB m-file software codes. In this way, the overall framework is tested and experiment outcomes are analyzed. According to our experiment outcomes, it is concluded that our closed loop cache aware adaptive scheduling framework successfully drives co-runner cache dependent thread instruction count to co-runner independent instruction count with an error margin up to 25% in case cache is highly utilized. In addition, thread cache access pattern is also estimated with 75% accuracy.

A constrained recursive least squares algorithm for adaptive combination of multiple models

In this paper, we develop a novel constrained recursive least squares algorithm for adaptively combining a set of given multiple models. With data available in an online fashion, the linear combination coefficients of submodels are adapted via the proposed algorithm.We propose to minimize the mean square error with a forgetting factor, and apply the sum to one constraint to the combination parameters. Moreover an l1-norm constraint to the combination parameters is also applied with the aim to achieve sparsity of multiple models so that only a subset of models may be selected into the final model. Then a weighted l2-norm is applied as an approximation to the l1-norm term. As such at each time step, a closed solution of the model combination parameters is available. The contribution of this paper is to derive the proposed constrained recursive least squares algorithm that is computational efficient by exploiting matrix theory. The effectiveness of the approach has been demonstrated using both simulated and real time series examples.

Adaptive recursive control of a soil testing load frame

Recursive Learning Control (RLC) has the potential to significantly reduce the tracking error in many repetitive trajectory applications. This paper presents an application of RLC to a soil testing load frame where non-adaptive techniques struggle with the highly nonlinear nature of soil. The main purpose of the controller is to apply a sinusoidal force reference trajectory on a soil sample with a high degree of accuracy and repeatability. The controller uses a feedforward control structure, recursive least squares adaptation algorithm and RLC to compensate for periodic errors. Tracking error is reduced and stability is maintained across various soil sample responses.

Adaptive stepsize based on control theory for stochastic differential equations

The numerical solution of stochastic differential equations (SDEs) has been focused recently on the development of numerical methods with good stability and order properties. These numerical implementations have been made with fixed stepsize, but there are many situations when a fixed stepsize is not appropriate. In the numerical solution of ordinary differential equations, much work has been carried out on developing robust implementation techniques using variable stepsize. It has been necessary, in the deterministic case, to consider the "best" choice for an initial stepsize, as well as developing effective strategies for stepsize control-the same, of course, must be carried out in the stochastic case. In this paper, proportional integral (PI) control is applied to a variable stepsize implementation of an embedded pair of stochastic Runge-Kutta methods used to obtain numerical solutions of nonstiff SDEs. For stiff SDEs, the embedded pair of the balanced Milstein and balanced implicit method is implemented in variable stepsize mode using a predictive controller for the stepsize change. The extension of these stepsize controllers from a digital filter theory point of view via PI with derivative (PID) control will also be implemented. The implementations show the improvement in efficiency that can be attained when using these control theory approaches compared with the regular stepsize change strategy.

Sparse multioutput radial basis function network construction using combined locally regularised orthogonal least square and D-optimality experimental design

A construction algorithm for multioutput radial basis function (RBF) network modelling is introduced by combining a locally regularised orthogonal least squares (LROLS) model selection with a D-optimality experimental design. The proposed algorithm aims to achieve maximised model robustness and sparsity via two effective and complementary approaches. The LROLS method alone is capable of producing a very parsimonious RBF network model with excellent generalisation performance. The D-optimality design criterion enhances the model efficiency and robustness. A further advantage of the combined approach is that the user only needs to specify a weighting for the D-optimality cost in the combined RBF model selecting criterion and the entire model construction procedure becomes automatic. The value of this weighting does not influence the model selection procedure critically and it can be chosen with ease from a wide range of values.

Modelagem e controle adaptativo de uma planta didática de nível com instrumentação industrial

The control, automation and optimization areas help to improve the processes used by industry. They contribute to a fast production line, improving the products quality and reducing the manufacturing costs. Didatic plants are good tools for research in these areas, providing a direct contact with some industrial equipaments. Given these capabilities, the main goal of this work is to model and control a didactic plant, which is a level and flow process control system with an industrial instrumentation. With a model it is possible to build a simulator for the plant that allows studies about its behaviour, without any of the real processes operational costs, like experiments with controllers. They can be tested several times before its application in a real process. Among the several types of controllers, it was used adaptive controllers, mainly the Direct Self-Tuning Regulators (DSTR) with Integral Action and the Gain Scheduling (GS). The DSTR was based on Pole-Placement design and use the Recursive Least Square to calculate the controller parameters. The characteristics of an adaptive system was very worth to guarantee a good performance when the controller was applied to the plant

Adaptive control system based on lineal control theory for the path-following problem of a car-like mobile robot

The objective of this paper is to design a path following control system for a car-like mobile robot using classical linear control techniques, so that it adapts on-line to varying conditions during the trajectory following task. The main advantages of the proposed control structure is that well known linear control theory can be applied in calculating the PID controllers to full control requirements, while at the same time it is exible to be applied in non-linear changing conditions of the path following task. For this purpose the Frenet frame kinematic model of the robot is linearised at a varying working point that is calculated as a function of the actual velocity, the path curvature and kinematic parameters of the robot, yielding a transfer function that varies during the trajectory. The proposed controller is formed by a combination of an adaptive PID and a feed-forward controller, which varies accordingly with the working conditions and compensates the non-linearity of the system. The good features and exibility of the proposed control structure have been demonstrated through realistic simulations that include both kinematics and dynamics of the car-like robot.

Adaptive stepsize based on control theory for stochastic differential equations

The numerical solution of stochastic differential equations (SDEs) has been focussed recently on the development of numerical methods with good stability and order properties. These numerical implementations have been made with fixed stepsize, but there are many situations when a fixed stepsize is not appropriate. In the numerical solution of ordinary differential equations, much work has been carried out on developing robust implementation techniques using variable stepsize. It has been necessary, in the deterministic case, to consider the best choice for an initial stepsize, as well as developing effective strategies for stepsize control-the same, of course, must be carried out in the stochastic case. In this paper, proportional integral (PI) control is applied to a variable stepsize implementation of an embedded pair of stochastic Runge-Kutta methods used to obtain numerical solutions of nonstiff SDEs. For stiff SDEs, the embedded pair of the balanced Milstein and balanced implicit method is implemented in variable stepsize mode using a predictive controller for the stepsize change. The extension of these stepsize controllers from a digital filter theory point of view via PI with derivative (PID) control will also be implemented. The implementations show the improvement in efficiency that can be attained when using these control theory approaches compared with the regular stepsize change strategy. (C) 2004 Elsevier B.V. All rights reserved.

Geometric Control Theory and its Application to Underwater Vehicles

This dissertation is based on theoretical study and experiments which extend geometric control theory to practical applications within the field of ocean engineering. We present a method for path planning and control design for underwater vehicles by use of the architecture of differential geometry. In addition to the theoretical design of the trajectory and control strategy, we demonstrate the effectiveness of the method via the implementation onto a test-bed autonomous underwater vehicle. Bridging the gap between theory and application is the ultimate goal of control theory. Major developments have occurred recently in the field of geometric control which narrow this gap and which promote research linking theory and application. In particular, Riemannian and affine differential geometry have proven to be a very effective approach to the modeling of mechanical systems such as underwater vehicles. In this framework, the application of a kinematic reduction allows us to calculate control strategies for fully and under-actuated vehicles via kinematic decoupled motion planning. However, this method has not yet been extended to account for external forces such as dissipative viscous drag and buoyancy induced potentials acting on a submerged vehicle. To fully bridge the gap between theory and application, this dissertation addresses the extension of this geometric control design method to include such forces. We incorporate the hydrodynamic drag experienced by the vehicle by modifying the Levi-Civita affine connection and demonstrate a method for the compensation of potential forces experienced during a prescribed motion. We present the design method for multiple different missions and include experimental results which validate both the extension of the theory and the ability to implement control strategies designed through the use of geometric techniques. By use of the extension presented in this dissertation, the underwater vehicle application successfully demonstrates the applicability of geometric methods to design implementable motion planning solutions for complex mechanical systems having equal or fewer input forces than available degrees of freedom. Thus, we provide another tool with which to further increase the autonomy of underwater vehicles.

Monotonous trend estimation using recursive prony analysis

This paper presents a solution to the problem of estimating the monotonous tendency of a slow-varying oscillating system. A recursive Prony Analysis (PA) scheme is developed which involves obtaining a dynamic model with parameters identified by implementing the forgetting factor recursive least square (FFRLS) method. A box threshold principle is proposed to separate the dominant components, which results in an accurate estimation of the trend of oscillating systems. Performance of the proposed PA is evaluated using real-time measurements when random noise and vibration effects are present. Moreover, the proposed method is used to estimate monotonous tendency of deck displacement to assist in a safe landing of an unmanned aerial vehicle (UAV). It is shown that the proposed method can estimate instantaneous mean deck satisfactorily, making it well suited for integration into ship-UAV approach and landing guidance systems.

Prediction of CO concentrations based on a hybrid Partial Least Square and Support Vector Machine model

Due to the health impacts caused by exposures to air pollutants in urban areas, monitoring and forecasting of air quality parameters have become popular as an important topic in atmospheric and environmental research today. The knowledge on the dynamics and complexity of air pollutants behavior has made artificial intelligence models as a useful tool for a more accurate pollutant concentration prediction. This paper focuses on an innovative method of daily air pollution prediction using combination of Support Vector Machine (SVM) as predictor and Partial Least Square (PLS) as a data selection tool based on the measured values of CO concentrations. The CO concentrations of Rey monitoring station in the south of Tehran, from Jan. 2007 to Feb. 2011, have been used to test the effectiveness of this method. The hourly CO concentrations have been predicted using the SVM and the hybrid PLS–SVM models. Similarly, daily CO concentrations have been predicted based on the aforementioned four years measured data. Results demonstrated that both models have good prediction ability; however the hybrid PLS–SVM has better accuracy. In the analysis presented in this paper, statistic estimators including relative mean errors, root mean squared errors and the mean absolute relative error have been employed to compare performances of the models. It has been concluded that the errors decrease after size reduction and coefficients of determination increase from 56 to 81% for SVM model to 65–85% for hybrid PLS–SVM model respectively. Also it was found that the hybrid PLS–SVM model required lower computational time than SVM model as expected, hence supporting the more accurate and faster prediction ability of hybrid PLS–SVM model.

Site characterization model using least-square support vector machine and relevance vector machine based on corrected SPT data (N-c)

Statistical learning algorithms provide a viable framework for geotechnical engineering modeling. This paper describes two statistical learning algorithms applied for site characterization modeling based on standard penetration test (SPT) data. More than 2700 field SPT values (N) have been collected from 766 boreholes spread over an area of 220 sqkm area in Bangalore. To get N corrected value (N,), N values have been corrected (Ne) for different parameters such as overburden stress, size of borehole, type of sampler, length of connecting rod, etc. In three-dimensional site characterization model, the function N-c=N-c (X, Y, Z), where X, Y and Z are the coordinates of a point corresponding to N, value, is to be approximated in which N, value at any half-space point in Bangalore can be determined. The first algorithm uses least-square support vector machine (LSSVM), which is related to aridge regression type of support vector machine. The second algorithm uses relevance vector machine (RVM), which combines the strengths of kernel-based methods and Bayesian theory to establish the relationships between a set of input vectors and a desired output. The paper also presents the comparative study between the developed LSSVM and RVM model for site characterization. Copyright (C) 2009 John Wiley & Sons,Ltd.

Engine modelling for diesel smoke feedback control system design

A novel smoke sensor was used to measure the smoke response to the fuel rack on a diesel engine. The conventional modelling methods used for engine control were investigated. The synchronization technique and Recursive Least Square method were applied to engine modelling and two models for controller design were derived.

Hard real-time multiprocessor scheduling resilient to core failures

Presented at 21st IEEE International Conference on Embedded and Real-Time Computing Systems and Applications (RTCSA 2015). 19 to 21, Aug, 2015, pp 122-131. Hong Kong, China.