Differential Migration: Sensitivity analysis and comparison study

Abstract

The contribution treats properties of a new evolutionary algorithm -Differential Migration, and provides a comparison with other algorithms of this type. Differential Migration is tested with a standard artificial neural network benchmark and standard test functions for performance comparison. Sensitivity analysis is conducted in order to specify the optimal parameters and their influence to the algorithm performance. SOMA (Self-Organizing Migration Algorithm) and Differential Evolution are used as a reference, and the results are compared with Differential Migration.

Temperature Control in Air-heating Set via Direct Search Method and Structured Singular Value

“…As the cost function defined by (24) as well as the upper bound has more than one local minimum an algorithm for global optimization is desirable. In this contribution, Differential Migration [2] was used for the optimization with high efficiency in finding the global minimum. Differential Migration is an evolutionary algorithm based on migration of individuals in the space of tuning parameters giving significantly higher robustness (in the sense of ability of finding the global minimum) than other algorithms of this class.…”

“…The algebraic μ-synthesis ( [3], [4], [5], [6], [7]) presented in this contribution overcomes both the approximation of the scaling matrices D, D -1 and the impossibility of integrating behaviour of the performance weighting function. The controller is designed through the algebraic pole placement principle applied to the nominal plant and the position of the nominal closed-loop poles is tuned through an evolutionary [2] algorithm with evaluation of the upper bound for . The problem of instability of performance weighting function is treated by setting the nominal closed-loop poles to the real axis in the left half-plane.…”

Robust control design toolbox for Time Delay Systems with Parametric Uncertainties

“…The poles of the nominal closed loop are tuned via direct search methodsDifferential Migration [9] and Nelder-Mead simplex method handling the problem of multimodality of the cost function. This method solves the problem of impossibility of usage of *This work was supported by the the weights with poles on imaginary axis and convergence to a global or even local minimum leading to non-optimality of the resulting controller in the D-K iteration [22].…”

Simple Robust Controllers for Delayed Systems

“…Evolutionary optimization by Differential Migration (DM, see Dlapa, 2009) (27) Both controllers satisfy condition (14) (see Fig. 9).…”