• Title/Summary/Keyword: weighted function

Search Result 749, Processing Time 0.023 seconds

A Novel Approach on $H_{\infty}$-LTR Controller Design ($H_{\infty}$-LTR 제어기 설계의 새로운 접근방법)

  • Lhee, Chin-Gook;Park, Jae-Sam;Ahn, Hyun-Sik;Kim, Do-Hyun
    • Journal of the Korean Institute of Telematics and Electronics S
    • /
    • v.36S no.2
    • /
    • pp.38-45
    • /
    • 1999
  • In this paper, A novel approach on $H_{\infty}-LTR$ design scheme is presented. The proposed scheme provides a design toll which can trade-off the recovery error against the control input. In the first stage, Kalman filter is designed to shape the loop to satisfy the required performance specifications. The designed Kalman filter, together with the plant transfer function, is used as a target transfer function. In the second stage, sensitivity function weighted $H_{\infty}-LTR$suboptimal LTR is designed to recover the target loop transfer function. Simulation results of LQG/LTR, $H_{\infty}-LTR$are compared to demonstrate the good property of the proposed scheme.

  • PDF

Function Approximation for accelerating learning speed in Reinforcement Learning (강화학습의 학습 가속을 위한 함수 근사 방법)

  • Lee, Young-Ah;Chung, Tae-Choong
    • Journal of the Korean Institute of Intelligent Systems
    • /
    • v.13 no.6
    • /
    • pp.635-642
    • /
    • 2003
  • Reinforcement learning got successful results in a lot of applications such as control and scheduling. Various function approximation methods have been studied in order to improve the learning speed and to solve the shortage of storage in the standard reinforcement learning algorithm of Q-Learning. Most function approximation methods remove some special quality of reinforcement learning and need prior knowledge and preprocessing. Fuzzy Q-Learning needs preprocessing to define fuzzy variables and Local Weighted Regression uses training examples. In this paper, we propose a function approximation method, Fuzzy Q-Map that is based on on-line fuzzy clustering. Fuzzy Q-Map classifies a query state and predicts a suitable action according to the membership degree. We applied the Fuzzy Q-Map, CMAC and LWR to the mountain car problem. Fuzzy Q-Map reached the optimal prediction rate faster than CMAC and the lower prediction rate was seen than LWR that uses training example.

Image quality assessment of color LCD monitors by polychromatic modulation transfer function (다색광전달함수를 사용한 컬러 LCD 모니터의 광학적 상평가법)

  • Song, Jong-Sup;Jo, Jae-Heung;Hong, Sung-Mok;Lee, Yun-Woo;Yang, Ho-Soon;Cho, Hyun-Mo;Lee, In-Won
    • Korean Journal of Optics and Photonics
    • /
    • v.16 no.1
    • /
    • pp.63-70
    • /
    • 2005
  • We propose a method for evaluating the image quality of color liquid crystal display(LCD) monitors by using the polychromatic modulation transfer function(PMTF), which is calculated from the modulation transfer function(MTF) weighted by the overall color response of the system including the test LCD monitor. We confirm that experimental results using the PMTF agree well with simulated results of the PMTF of a color LCD monitor by using three bar targets with different amplitudes and three elementary colors such as red(R), green(G), and blue(B). As a results, we should choose the PMTF instead of the white color MTF or monochromatic MTF in order to evaluate correctly the image quality of color LCD monitors.

Weight Adjustment Scheme Based on Hop Count in Q-routing for Software Defined Networks-enabled Wireless Sensor Networks

  • Godfrey, Daniel;Jang, Jinsoo;Kim, Ki-Il
    • Journal of information and communication convergence engineering
    • /
    • v.20 no.1
    • /
    • pp.22-30
    • /
    • 2022
  • The reinforcement learning algorithm has proven its potential in solving sequential decision-making problems under uncertainties, such as finding paths to route data packets in wireless sensor networks. With reinforcement learning, the computation of the optimum path requires careful definition of the so-called reward function, which is defined as a linear function that aggregates multiple objective functions into a single objective to compute a numerical value (reward) to be maximized. In a typical defined linear reward function, the multiple objectives to be optimized are integrated in the form of a weighted sum with fixed weighting factors for all learning agents. This study proposes a reinforcement learning -based routing protocol for wireless sensor network, where different learning agents prioritize different objective goals by assigning weighting factors to the aggregated objectives of the reward function. We assign appropriate weighting factors to the objectives in the reward function of a sensor node according to its hop-count distance to the sink node. We expect this approach to enhance the effectiveness of multi-objective reinforcement learning for wireless sensor networks with a balanced trade-off among competing parameters. Furthermore, we propose SDN (Software Defined Networks) architecture with multiple controllers for constant network monitoring to allow learning agents to adapt according to the dynamics of the network conditions. Simulation results show that our proposed scheme enhances the performance of wireless sensor network under varied conditions, such as the node density and traffic intensity, with a good trade-off among competing performance metrics.

T-S fuzzy PID control based on RCGAs for the automatic steering system of a ship (선박자동조타를 위한 RCGA기반 T-S 퍼지 PID 제어)

  • Yu-Soo LEE;Soon-Kyu HWANG;Jong-Kap AHN
    • Journal of the Korean Society of Fisheries and Ocean Technology
    • /
    • v.59 no.1
    • /
    • pp.44-54
    • /
    • 2023
  • In this study, the second-order Nomoto's nonlinear expansion model was implemented as a Tagaki-Sugeno fuzzy model based on the heading angular velocity to design the automatic steering system of a ship considering nonlinear elements. A Tagaki-Sugeno fuzzy PID controller was designed using the applied fuzzy membership functions from the Tagaki-Sugeno fuzzy model. The linear models and fuzzy membership functions of each operating point of a given nonlinear expansion model were simultaneously tuned using a genetic algorithm. It was confirmed that the implemented Tagaki-Sugeno fuzzy model could accurately describe the given nonlinear expansion model through the Zig-Zag experiment. The optimal parameters of the sub-PID controller for each operating point of the Tagaki-Sugeno fuzzy model were searched using a genetic algorithm. The evaluation function for searching the optimal parameters considered the route extension due to course deviation and the resistance component of the ship by steering. By adding a penalty function to the evaluation function, the performance of the automatic steering system of the ship could be evaluated to track the set course without overshooting when changing the course. It was confirmed that the sub-PID controller for each operating point followed the set course to minimize the evaluation function without overshoot when changing the course. The outputs of the tuned sub-PID controllers were combined in a weighted average method using the membership functions of the Tagaki-Sugeno fuzzy model. The proposed Tagaki-Sugeno fuzzy PID controller was applied to the second-order Nomoto's nonlinear expansion model. As a result of examining the transient response characteristics for the set course change, it was confirmed that the set course tracking was satisfactorily performed.

Pole Placement Method to Move a Equal Poles with Jordan Block to Two Real Poles Using LQ Control and Pole's Moving-Range (LQ 제어와 근의 이동범위를 이용한 조단 블록을 갖는 중근을 두 실근으로 이동시키는 극배치 방법)

  • Park, Minho
    • Journal of the Korea Academia-Industrial cooperation Society
    • /
    • v.19 no.2
    • /
    • pp.608-616
    • /
    • 2018
  • If a general nonlinear system is linearized by the successive multiplication of the 1st and 2nd order systems, then there are four types of poles in this linearized system: the pole of the 1st order system and the equal poles, two distinct real poles, and complex conjugate pair of poles of the 2nd order system. Linear Quadratic (LQ) control is a method of designing a control law that minimizes the quadratic performance index. It has the advantage of ensuring the stability of the system and the pole placement of the root of the system by weighted matrix adjustment. LQ control by the weighted matrix can move the position of the pole of the system arbitrarily, but it is difficult to set the weighting matrix by the trial and error method. This problem can be solved using the characteristic equations of the Hamiltonian system, and if the control weighting matrix is a symmetric matrix of constants, it is possible to move several poles of the system to the desired closed loop poles by applying the control law repeatedly. The paper presents a method of calculating the state weighting matrix and the control law for moving the equal poles with Jordan blocks to two real poles using the characteristic equation of the Hamiltonian system. We express this characteristic equation with a state weighting matrix by means of a trigonometric function, and we derive the relation function (${\rho},\;{\theta}$) between the equal poles and the state weighting matrix under the condition that the two real poles are the roots of the characteristic equation. Then, we obtain the moving-range of the two real poles under the condition that the state weighting matrix becomes a positive semi-finite matrix. We calculate the state weighting matrix and the control law by substituting the two real roots selected in the moving-range into the relational function. As an example, we apply the proposed method to a simple example 3rd order system.

WEAKTYPE $L^1(R^n)$-ESTIMATE FOR CRETAIN MAXIMAL OPERATORS

  • Kim, Yong-Cheol
    • Journal of the Korean Mathematical Society
    • /
    • v.34 no.4
    • /
    • pp.1029-1036
    • /
    • 1997
  • Let ${A_t)}_{t>0}$ be a dilation group given by $A_t = exp(-P log t)$, where P is a real $n \times n$ matrix whose eigenvalues has strictly positive real part. Let $\nu$ be the trace of P and $P^*$ denote the adjoint of pp. Suppose that $K$ is a function defined on $R^n$ such that $$\mid$K(x)$\mid$ \leq k($\mid$x$\mid$_Q)$ for a bounded and decreasing function $k(t) on R_+$ satisfying $k \diamond $\mid$\cdot$\mid$_Q \in \cup_{\varepsilon >0}L^1((1 + $\mid$x$\mid$)^\varepsilon dx)$ where $Q = \int_{0}^{\infty} exp(-tP^*) exp(-tP)$ dt and the norm $$\mid$\cdot$\mid$_Q$ stands for $$\mid$x$\mid$_Q = \sqrt{}, x \in R^n$. For $f \in L^1(R^n)$, define $mf(x) = sup_{t>0}$\mid$K_t * f(x)$\mid$$ where $K_t(X) = t^{-\nu}K(A_{1/t}^* x)$. Then we show that $m$ is a bounded operator of $L^1(R^n) into L^{1, \infty}(R^n)$.

  • PDF

Daily Electric Load Forecasting Based on RBF Neural Network Models

  • Hwang, Heesoo
    • International Journal of Fuzzy Logic and Intelligent Systems
    • /
    • v.13 no.1
    • /
    • pp.39-49
    • /
    • 2013
  • This paper presents a method of improving the performance of a day-ahead 24-h load curve and peak load forecasting. The next-day load curve is forecasted using radial basis function (RBF) neural network models built using the best design parameters. To improve the forecasting accuracy, the load curve forecasted using the RBF network models is corrected by the weighted sum of both the error of the current prediction and the change in the errors between the current and the previous prediction. The optimal weights (called "gains" in the error correction) are identified by differential evolution. The peak load forecasted by the RBF network models is also corrected by combining the load curve outputs of the RBF models by linear addition with 24 coefficients. The optimal coefficients for reducing both the forecasting mean absolute percent error (MAPE) and the sum of errors are also identified using differential evolution. The proposed models are trained and tested using four years of hourly load data obtained from the Korea Power Exchange. Simulation results reveal satisfactory forecasts: 1.230% MAPE for daily peak load and 1.128% MAPE for daily load curve.

Shape Optimization of a Rotating Two-Pass Duct with a Guide Vane in the Turning Region (회전하는 냉각유로의 곡관부에 부착된 가이드 베인의 형상 최적설계)

  • Moon, Mi-Ae;Kim, Kwang-Yong
    • The KSFM Journal of Fluid Machinery
    • /
    • v.14 no.1
    • /
    • pp.66-76
    • /
    • 2011
  • The heat transfer and pressure loss characteristics of a rotating two-pass channel with a guide vane in the turning region have been studied using three-dimensional Reynolds-averaged Navier-Stokes (RANS) analysis, and the shape of the guide vane has been optimized using surrogate modeling optimization technique. For the optimization, thickness, location and angle of the guide vanes have been selected as design variables. The objective function has been defined as a linear combination of the heat transfer and the friction loss related terms with a weighting factor. Latin hypercube sampling has been applied to determine the design points as design of experiments. A weighted-average surrogate model, PBA has been used as the surrogate model. The guide vane in the turning region does not influence the heat transfer in the first passage upstream of the turning region, but enhances largely the heat transfer in the turning region and the second passage. In an example of the optimization, the objective function has been increased by 13.6%.

EXPLORING THE FUEL ECONOMY POTENTIAL OF ISG HYBRID ELECTRIC VEHICLES THROUGH DYNAMIC PROGRAMMING

  • Ao, G.Q.;Qiang, J.X.;Zhong, H.;Yang, L.;Zhuo, B.
    • International Journal of Automotive Technology
    • /
    • v.8 no.6
    • /
    • pp.781-790
    • /
    • 2007
  • Hybrid electric vehicles(HEV) combined with more than one power sources have great potential to improve fuel economy and reduce pollutant emissions. The Integrated Starter Generator(ISG) HEV researched in this paper is a two energy sources vehicle, with a conventional internal combustion engine(ICE) and an energy storage system(batteries). In order to investigate the potential of diesel engine hybrid electric vehicles in fuel economy improvement and emissions reduction, a Dynamic Programming(DP) based supervisory controller is developed to allocate the power requirement between ICE and batteries with the objective of minimizing a weighted cost function over given drive cycles. A fuel-economy-only case and a fuel & emissions case can be achieved by changing specific weighting factors. The simulation results of the fuel-economy-only case show that there is a 45.1% fuel saving potential for this ISG HEV compared to a conventional transit bus. The test results present a 39.6% improvement in fuel economy which validates the simulation results. Compared to the fuel-economy-only case, the fuel & emissions case further reduces the pollutant emissions at a cost of 3.2% and 4.5% of fuel consumption with respect to the simulation and test result respectively.