Udit Halder

A Cluster-Based Differential Evolution Algorithm With External Archive for Optimization in Dynamic Environments

This paper presents a Cluster-based Dynamic Differential Evolution with external Ar chive (CDDE_Ar) for global optimization in dynamic fitness landscape. The algorithm uses a multipopulation method where the entire population is partitioned into several clusters according to the spatial locations of the trial solutions. The clusters are evolved separately using a standard differential evolution algorithm. The number of clusters is an adaptive parameter, and its value is updated after a certain number of iterations. Accordingly, the total population is redistributed into a new number of clusters. In this way, a certain sharing of information occurs periodically during the optimization process. The performance of CDDE_Ar is compared with six state-of-the-art dynamic optimizers over the moving peaks benchmark problems and dynamic optimization problem (DOP) benchmarks generated with the generalized-dynamic-benchmark-generator system for the competition and special session on dynamic optimization held under the 2009 IEEE Congress on Evolutionary Computation. Experimental results indicate that CDDE_Ar can enjoy a statistically superior performance on a wide range of DOPs in comparison to some of the best known dynamic evolutionary optimizers.

Chaotic Dynamics in Social Foraging Swarms—An Analysis

This paper investigates the chaotic characteristics in the dynamics of an aggregating swarm model. The range of the parameters of the swarm model is determined for which chaos exists in the dynamics. The trajectories of the individuals are simulated, and the stable, limit cyclic, and chaotic behaviors are demonstrated. The existence of chaos in the swarm is determined by the maximum Lyapunov exponent. The computer simulation supports the results obtained by theoretical analysis.

Self adaptive cluster based and weed inspired differential evolution algorithm for real world optimization

In this paper we propose a Self Adaptive Cluster based and Weed Inspired Differential Evolution algorithm (SACWIDE), the total population is divided into several clusters based on the positions of the individuals and the cluster number is dynamically changed by the suitable learning strategy during evolution. Here we incorporate a modified version of the Invasive Weed Optimization (IWO) algorithm as a local search technique. The algorithm strategically determines whether a particular cluster will perform Differential Evolution (DE) or the IWO algorithm (modified). The number of clusters in a particular iteration is set by the algorithm itself self-adaptively. The performance of SACWIDE is reported on the set of 22 benchmark problems of CEC-2011.

An Informative Differential Evolution Algorithm with Self Adaptive Re-Clustering Technique for the Optimization of Phased Antenna Array

In this paper, we propose a new algorithm called An Informative Difierential Evolution with Self Adaptive Re-clustering Technique to flnd the amplitude-phase excitation of a linear phased array to have the desired far fleld pattern. Here we consider three problems for three difierent far fleld patterns and each problem is optimized with this algorithm. This algorithm has a proper balancing of exploration and exploitation power which is achieved with the help of information exchange among the subpopulations. We also used an elitist local search algorithm for the flne tuning at the suspected optimal position, and that helps us from the unnecessary wastage of Function Evaluations (FEs).

An informative differential evolution with self adaptive re-clustering technique

We propose an informative Differential Evolution (DE) algorithm where the information gained by the individuals of a cluster will be exchanged after a certain number of iterations called refreshing gap. The DE is empowered with a clustering technique to improve its efficiency over multimodal landscapes. During evolution, self-adaptive behaviour helps in re-clustering. With the better explorative power of the proposed algorithm we have used a new local search technique for fine tuning near a suspected optimal position. The performance of the proposed algorithm is evaluated over 25 benchmark functions and compared with existing algorithms.

Self-adaptive cluster-based differential evolution with an external archive for dynamic optimization problems

In this paper we propose a self adaptive cluster based Differential Evolution (DE) algorithm to solve the Dynamic Optimization Problems (DOPs). We have enhanced the classical DE to perform better in dynamic environments by a powerful clustering technique. During evolution, the information gained by the particles of different clusters is exchanged by a self adaptive strategy. The information exchange is done by re-clustering, and the cluster number is updated adaptively throughout the optimization process. To detect the environment change a test particle is used. Moreover, to adapt the population in new environment an External Archive is also used. The performance of SACDEEA is evaluated on GDBG benchmark problems and compared with other existing algorithms.

Convergence and boundary estimation of the particle dynamics in generalized particle swarm optimization

Concept of the particle swarms emerged from a simulation of the collective behavior of social creatures and gradually evolved into a powerful global optimization technique, now well-known as the Particle Swarm Optimization (PSO). A vast amount of analytical studies on various aspects of the PSO dynamics like stability, convergence, explorative power, sampling distribution and so on can be found in the literature. The boundary of the swarm is still as a challenging research interest. The upper boundary restricts the swarm members within a sub-region of the whole search space. Higher the upper boundary, higher is the diversity. This paper investigates mainly the diversity of the swarm in terms of the upper boundary of the swarm.

A frequency domain approach to silhouette based gait recognition

In this paper, we propose a frequency domain based model-free gait recognition approach from silhouette inputs using Fourier Transform. Gait sequences are first converted into frequency domain using Fourier transform. Information content of the frequency components are analysed next to determine the number of effective frequencies which can help in the recognition process. These principal frequencies are treated separately to obtain scores based on the correlation coefficient between the gallery and the probe images. The individual scores are fused in the last stage to obtain the final score. The proposed approach is compared with other state-of-the-art model-free gait recognition algorithms. Experimental results on the USF HumanID database clearly indicate the supremacy of our technique.

On the non linear dynamics of the global best particle in particle swarm optimization

The dynamics of Particle swarm optimization has been developed from the collective behavior of the social creatures like fish schooling and gradually has become a powerful global optimization technique. In this paper we do the analysis on a continuous variant of PSO. The non linear dynamics of the global best particle is studied here and the exponential convergence is ensured. The effects of the different control parameters on the convergence of the global best particle are also studied.

Performance of informative differential evolution algorithm with self adaptive re-clustering technique on the problems of electromagnetism: the linear array antenna synthesis

In this paper we are going to investigate the properties of Informative Differential Evolution with Self Adaptive Re-clustering technique (IDE_SR) algorithm on the problems of electromagnetic domain. Problems of electromagnetism domain generally exhibit multimodal characteristic as well as the high dimensionality and non-smooth attributes. IDE_SR is a modified Differential Evolution (DE) to overcome the problems of multimodality and complexity of the problem under consideration. In this paper we describe the salient features of IDE_SR and compare the result with other state-of-art algorithms.