scholarly journals Difference-Based Mutation Operation for Neuroevolution of Augmented Topologies

Algorithms ◽  
2021 ◽  
Vol 14 (5) ◽  
pp. 127
Author(s):  
Vladimir Stanovov ◽  
Shakhnaz Akhmedova ◽  
Eugene Semenkin
Keyword(s):  
Inverted Pendulum ◽  
Control Algorithms ◽  
Classification Problems ◽  
Optimal Weights ◽  
Mutation Operators ◽  
Search Operation ◽  
The Difference ◽  
Parameter Values ◽  
Rotary Inverted Pendulum ◽  
Classification Quality

In this paper, a novel search operation is proposed for the neuroevolution of augmented topologies, namely the difference-based mutation. This operator uses the differences between individuals in the population to perform more efficient search for optimal weights and structure of the model. The difference is determined according to the innovation numbers assigned to each node and connection, allowing tracking the changes. The implemented neuroevolution algorithm allows backward connections and loops in the topology, and uses a set of mutation operators, including connections merging and deletion. The algorithm is tested on a set of classification problems and the rotary inverted pendulum control problem. The comparison is performed between the basic approach and modified versions. The sensitivity to parameter values is examined. The experimental results prove that the newly developed operator delivers significant improvements to the classification quality in several cases, and allow finding better control algorithms.

scholarly journals Fractional Norms and Quasinorms Do Not Help to Overcome the Curse of Dimensionality

Entropy ◽  
2020 ◽  
Vol 22 (10) ◽  
pp. 1105
Author(s):  
Evgeny M. Mirkes ◽  
Jeza Allohibi ◽  
Alexander Gorban
Keyword(s):  
Machine Learning ◽  
Space Dimension ◽  
Curse Of Dimensionality ◽  
Manhattan Distance ◽  
Classification Problems ◽  
Machine Learning Methods ◽  
Relative Contrast ◽  
The Difference ◽  
Definition Of ◽  
Classification Quality

The curse of dimensionality causes the well-known and widely discussed problems for machine learning methods. There is a hypothesis that using the Manhattan distance and even fractional lp quasinorms (for p less than 1) can help to overcome the curse of dimensionality in classification problems. In this study, we systematically test this hypothesis. It is illustrated that fractional quasinorms have a greater relative contrast and coefficient of variation than the Euclidean norm l2, but it is shown that this difference decays with increasing space dimension. It has been demonstrated that the concentration of distances shows qualitatively the same behaviour for all tested norms and quasinorms. It is shown that a greater relative contrast does not mean a better classification quality. It was revealed that for different databases the best (worst) performance was achieved under different norms (quasinorms). A systematic comparison shows that the difference in the performance of kNN classifiers for lp at p = 0.5, 1, and 2 is statistically insignificant. Analysis of curse and blessing of dimensionality requires careful definition of data dimensionality that rarely coincides with the number of attributes. We systematically examined several intrinsic dimensions of the data.

Current development on using Rotary Inverted Pendulum as a benchmark for testing linear and nonlinear control algorithms

2019 ◽  
Vol 116 ◽  
pp. 347-369 ◽  
Author(s):  
Mukhtar Fatihu Hamza ◽  
Hwa Jen Yap ◽  
Imtiaz Ahmed Choudhury ◽  
Abdulbasid Ismail Isa ◽  
Aminu Yahaya Zimit ◽  
...  
Keyword(s):  
Nonlinear Control ◽  
Inverted Pendulum ◽  
Control Algorithms ◽  
Current Development ◽  
Linear And Nonlinear ◽  
Rotary Inverted Pendulum ◽  
Testing Linear

scholarly journals A real-time HIL control system on rotary inverted pendulum hardware platform based on double deep Q-network

2021 ◽  
Vol 54 (3-4) ◽  
pp. 417-428
Author(s):  
Yanyan Dai ◽  
KiDong Lee ◽  
SukGyu Lee
Keyword(s):  
Control System ◽  
Reinforcement Learning ◽  
Inverted Pendulum ◽  
Learning Algorithm ◽  
Deep Understanding ◽  
Control Engineering ◽  
Experience Replay ◽  
Real Hardware ◽  
Rotary Inverted Pendulum ◽  
Reinforcement Learning Algorithm

For real applications, rotary inverted pendulum systems have been known as the basic model in nonlinear control systems. If researchers have no deep understanding of control, it is difficult to control a rotary inverted pendulum platform using classic control engineering models, as shown in section 2.1. Therefore, without classic control theory, this paper controls the platform by training and testing reinforcement learning algorithm. Many recent achievements in reinforcement learning (RL) have become possible, but there is a lack of research to quickly test high-frequency RL algorithms using real hardware environment. In this paper, we propose a real-time Hardware-in-the-loop (HIL) control system to train and test the deep reinforcement learning algorithm from simulation to real hardware implementation. The Double Deep Q-Network (DDQN) with prioritized experience replay reinforcement learning algorithm, without a deep understanding of classical control engineering, is used to implement the agent. For the real experiment, to swing up the rotary inverted pendulum and make the pendulum smoothly move, we define 21 actions to swing up and balance the pendulum. Comparing Deep Q-Network (DQN), the DDQN with prioritized experience replay algorithm removes the overestimate of Q value and decreases the training time. Finally, this paper shows the experiment results with comparisons of classic control theory and different reinforcement learning algorithms.

scholarly journals Archie's law – a reappraisal

Solid Earth ◽  
2016 ◽  
Vol 7 (4) ◽  
pp. 1157-1169 ◽  
Author(s):  
Paul W. J. Glover
Keyword(s):  
Systematic Errors ◽  
Apparent Paradox ◽  
Archie’S Law ◽  
A Value ◽  
The Difference ◽  
Reservoir Parameter ◽  
Parameter Values ◽  
Electrical Data ◽  
Extra Parameter ◽  
Better Than

Abstract. When scientists apply Archie's first law they often include an extra parameter a, which was introduced about 10 years after the equation's first publication by Winsauer et al. (1952), and which is sometimes called the “tortuosity” or “lithology” parameter. This parameter is not, however, theoretically justified. Paradoxically, the Winsauer et al. (1952) form of Archie's law often performs better than the original, more theoretically correct version. The difference in the cementation exponent calculated from these two forms of Archie's law is important, and can lead to a misestimation of reserves by at least 20 % for typical reservoir parameter values. We have examined the apparent paradox, and conclude that while the theoretical form of the law is correct, the data that we have been analysing with Archie's law have been in error. There are at least three types of systematic error that are present in most measurements: (i) a porosity error, (ii) a pore fluid salinity error, and (iii) a temperature error. Each of these systematic errors is sufficient to ensure that a non-unity value of the parameter a is required in order to fit the electrical data well. Fortunately, the inclusion of this parameter in the fit has compensated for the presence of the systematic errors in the electrical and porosity data, leading to a value of cementation exponent that is correct. The exceptions are those cementation exponents that have been calculated for individual core plugs. We make a number of recommendations for reducing the systematic errors that contribute to the problem and suggest that the value of the parameter a may now be used as an indication of data quality.

scholarly journals Breeding schemes in reindeer husbandry

Rangifer ◽  
2003 ◽  
Vol 23 (2) ◽  
pp. 45 ◽  
Author(s):  
Lars Rönnegård ◽  
J. A. Woolliams ◽  
Öje Danell
Keyword(s):  
Genetic Gain ◽  
Mass Selection ◽  
Effective Population ◽  
Selection Programme ◽  
Breeding Schemes ◽  
The Difference ◽  
Genetic Contributions ◽  
Parameter Values ◽  
Open Nucleus

The objective of the paper was to investigate annual genetic gain from selection (G), and the influence of selection on the inbreeding effective population size (Ne), for different possible breeding schemes within a reindeer herding district. The breeding schemes were analysed for different proportions of the population within a herding district included in the selection programme. Two different breeding schemes were analysed: an open nucleus scheme where males mix and mate between owner flocks, and a closed nucleus scheme where the males in non-selected owner flocks are culled to maximise G in the whole population. The theory of expected long-term genetic contributions was used and maternal effects were included in the analyses. Realistic parameter values were used for the population, modelled with 5000 reindeer in the population and a sex ratio of 14 adult females per male. The standard deviation of calf weights was 4.1 kg. Four different situations were explored and the results showed: 1. When the population was randomly culled, Ne equalled 2400. 2. When the whole population was selected on calf weights, Ne equalled 1700 and the total annual genetic gain (direct + maternal) in calf weight was 0.42 kg. 3. For the open nucleus scheme, G increased monotonically from 0 to 0.42 kg as the proportion of the population included in the selection programme increased from 0 to 1.0, and Ne decreased correspondingly from 2400 to 1700. 4. In the closed nucleus scheme the lowest value of Ne was 1300. For a given proportion of the population included in the selection programme, the difference in G between a closed nucleus scheme and an open one was up to 0.13 kg. We conclude that for mass selection based on calf weights in herding districts with 2000 animals or more, there are no risks of inbreeding effects caused by selection.

scholarly journals Analytical and Experimental Analyses of Nonlinear Vibrations in a Rotary Inverted Pendulum

2021 ◽  
Author(s):  
Mohammadjavad Rahimi dolatabad ◽  
Abdolreza Pasharavesh ◽  
Amir Ali Akbar Khayyat
Keyword(s):  
Analytical Solution ◽  
Numerical Solution ◽  
Closed Loop ◽  
Inverted Pendulum ◽  
Multiple Scales ◽  
Nonlinear Oscillations ◽  
Wide Frequency Range ◽  
Perturbative Method ◽  
Full State Feedback ◽  
Rotary Inverted Pendulum

Abstract Gaining insight into possible vibratory responses of dynamical systems around their stable equilibria is an essential step, which must be taken before their design and application. The results of such a study can significantly help prevent instability in closed-loop stabilized systems through avoiding the excitation of the system in the neighborhood of its resonance. This paper investigates nonlinear oscillations of a Rotary Inverted Pendulum (RIP) with a full-state feedback controller. Lagrange’s equations are employed to derive an accurate 2-DoF mathematical model, whose parameter values are extracted by both the measurement and 3D modeling of the real system components. Although the governing equations of a 2-DoF nonlinear system are difficult to solve, performing an analytical solution is of great importance, mostly because, compared to the numerical solution, the analytical solution can function as an accurate pattern. Additionally, the analytical solution is generally more appealing to engineers because their computational costs are less than those of the numerical solution. In this study, the perturbative method of multiple scales is used to obtain an analytical solution to the coupled nonlinear motion equations of the closed-loop system. Moreover, the parameters of the controller are determined, using the results of this solution. The findings reveal the existence of hardening- and softening-type resonances at the first and second vibrational modes, respectively. This led to a wide frequency range with moderately large-amplitude vibrations, which must be avoided when adjusting a time-varying set-point for the system. The analytical results of the nonlinear vibration of the RIP are verified by experimental measurements, and a very good agreement is observed between the results of both approaches.

scholarly journals Switching Stabilization for Nonlinear Networked Control Systems with Delays and Packet Losses

2020 ◽  
Vol 49 (2) ◽  
pp. 302-316
Author(s):  
Qi Zhang ◽  
Bin Liu
Keyword(s):  
Inverted Pendulum ◽  
Fuzzy Controller ◽  
Networked Control Systems ◽  
Fuzzy Model ◽  
Average Dwell Time ◽  
Switched System ◽  
Packet Dropout ◽  
Packet Losses ◽  
Nonlinear Networked Control Systems ◽  
Rotary Inverted Pendulum

This paper studies the stabilization problem for nonlinear NCSs(NNCSs) with bilateral network-induced random delay and packet dropout. T-S fuzzy model is employed to represent the nonlinear controlled plant. Based on the T-S model, a discrete-time fuzzy switched system model with uncertain parameters is established by means of the uncertain method and switching system method. Furthermore, the exponential stability condition for the state of the fuzzy switched system is obtained by using the combination of slow switching model-dependent average dwell time (MDADT) method and fast switching MDADT method. Finally, a series of rotary inverted pendulum experiments are provided to illustrates the effectiveness of the proposed method and prove that the proposed fuzzy controller based on T-S fuzzy model can balance the rotary inverted pendulum in a greater state range rather than the linear controller based on linearization

Sign in / Sign up

Export Citation Format

Share Document