Related Experiment Video
Updated: Jun 20, 2026
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
Locating and characterizing the stationary points of the extended Rosenbrock function.
1Department of Mechanical and Aeronautical Engineering, University of Pretoria, Pretoria, 0002, South Africa. schalk.kok@up.ac.za
Evolutionary Computation
|August 28, 2009
Summary
Analyzing the extended Rosenbrock function reveals its first variant has one global minimum. The second variant, however, presents numerous stationary points, including saddle points, making optimization challenging.
Area of Science:
- Numerical Analysis
- Optimization Theory
Background:
- The extended Rosenbrock function is a common benchmark for optimization algorithms.
- Identifying stationary points (minima, maxima, saddle points) is crucial for understanding function behavior.
Purpose of the Study:
- To analyze stationary points of two variants of the extended Rosenbrock function.
- To evaluate the efficacy of a standard Newtonian method with multi-start for locating these points.
- To characterize saddle points using Hessian eigenvalues.
Main Methods:
- Analysis of two variants of the extended Rosenbrock function.
- Application of a standard Newtonian method with multi-start.
- Computation of Hessian eigenvalues to classify stationary points.
- Monte Carlo simulations and particle swarm algorithm testing.
Main Results:
- The first variant has a single global minimum.
- The second variant exhibits numerous stationary points, especially in high dimensions.
- A previously proposed method is numerically intractable.
- Newtonian method with multi-start successfully locates stationary points.
- Many saddle points exist, with two having a single negative eigenvalue in high dimensions.
- A method to estimate the number of saddle points is proposed.
Conclusions:
- The second variant of the extended Rosenbrock function poses significant optimization challenges due to numerous saddle points.
- Standard optimization algorithms like uniform random search and particle swarm struggle to escape saddle points.
- The Newtonian method with multi-start is a viable approach for locating stationary points.
- Characterizing saddle points is essential for understanding optimization landscapes.
Related Concept Videos
01:19Plotting and Calibrating the Root Locus
Root loci often diverge as system poles shift from the real axis to the complex plane. Key points in this transition are the breakaway and break-in points, indicating where the root locus leaves and reenters the real axis. The branches of the root locus form an angle of 180/n degrees with the real axis, where n is the number of branches at a breakaway or break-in point.
The maximum gain occurs at the breakaway points between open-loop poles on the real axis, while the minimum gain is observed...
The maximum gain occurs at the breakaway points between open-loop poles on the real axis, while the minimum gain is observed...
01:15Construction of Root Locus
The construction of a root locus involves several key steps to analyze and visualize the behavior of a system's poles with varying gain. The number of branches in the root locus equals the number of closed-loop poles and is symmetrical about the real axis.
For positive gain values, the root locus exists on the real axis to the left of an odd number of finite open-loop poles or zeros. The root locus starts at the open-loop poles and traces the paths of the closed-loop poles as the gain increases.
For positive gain values, the root locus exists on the real axis to the left of an odd number of finite open-loop poles or zeros. The root locus starts at the open-loop poles and traces the paths of the closed-loop poles as the gain increases.
01:19Routh-Hurwitz Criterion II
In the application of the Routh-Hurwitz criterion, two specific scenarios can arise that complicate stability analysis.
The first scenario occurs when a singular zero appears in the first column of the Routh table. This situation creates a division by zero issues. To resolve this, a small positive or negative number, denoted as epsilon (∈), is substituted for the zero. The stability analysis proceeds by assuming a sign for ∈. If ∈ is positive, any sign change in the first column of the Routh...
The first scenario occurs when a singular zero appears in the first column of the Routh table. This situation creates a division by zero issues. To resolve this, a small positive or negative number, denoted as epsilon (∈), is substituted for the zero. The stability analysis proceeds by assuming a sign for ∈. If ∈ is positive, any sign change in the first column of the Routh...
01:05Properties of the Root Locus
The root locus method is an invaluable tool for analyzing higher-order systems without needing to factor the denominator of the transfer function. A pole of the system is identified when the characteristic polynomial in the transfer function's denominator equals zero.
To determine if a point lies on the root locus, the criterion involves the sum of angles contributed by all poles and zeros to that point. Specifically, this sum must be an odd multiple of 180 degrees. The gain at any point on the...
To determine if a point lies on the root locus, the criterion involves the sum of angles contributed by all poles and zeros to that point. Specifically, this sum must be an odd multiple of 180 degrees. The gain at any point on the...
01:09Rolle’s Theorem
Rolle’s Theorem states that if a real-valued function is continuous on a closed interval, differentiable on the open interval, and takes equal values at both endpoints, then there is at least one point within the open interval where the derivative of the function is zero.Rolle’s Theorem describes an important property of differentiable functions, this theorem applies to a real-valued function defined on a closed interval, provided three specific conditions are met. First, the function must be...
01:20Region of Convergence of Laplace Tarnsform
The Region of Convergence (ROC) is a fundamental concept in signal processing and system analysis, particularly associated with the Laplace transform. The ROC represents an area in the complex plane where the Laplace transform of a given signal converges, determining the transform's applicability and utility.
Consider a decaying exponential signal that begins at a specific time. When deriving its Laplace transform, the time-domain variable is replaced with a complex variable. This substitution...
Consider a decaying exponential signal that begins at a specific time. When deriving its Laplace transform, the time-domain variable is replaced with a complex variable. This substitution...