Premature Convergence Of Genetic Algorithm Research Articles

Multi-objective macrogeometry optimization of gears: Comparison between sequential quadratic programing and genetic algorithm

Deciding a gear pair’s macrogeometry parameters requires consideration of the factors such as cost of production, gear strength, and noise contributing parameters. A practical engineering problem is encountered when best possible gear geometry is needed for a fixed center distance to achieve the conflicting objectives. Further, constraints of practical utility on input parameters need to be taken into consideration for deciding the design space. This work represents comprehensive compilation of formulations of these often-conflicting objectives and constraints as non-linear functions and selection of an appropriate optimization technique to achieve a feasible design through a case study of a helical gear pair in high load carrying capacity mesh. For comparison of the quality of results and computational efficiency between gradient based and gradient free algorithms, Sequential Quadratic Programming and Genetic algorithm were used. The differences in the formulations of the objective functions for two algorithms were analyzed and addressed. Quadratic mutation rates and random crossovers were used to prevent premature convergence of Genetic Algorithm. Significant possible improvement in the base design was observed with both the algorithms.

Inversion method of particle size distribution of milk fat based on improved MPGA.

Milk fat’s particle size and distribution not only affect product quality, but also have great impacts on food safety in the economy and society. Based on total light scattering method, this paper has studied the inversion method of particle size distribution under dependent mode condition by combining multi-population genetic algorithm (MPGA) with Tikhonov smooth function. It has minimized the influence from light-absorb medium to improve the inversion accuracy. The approach introduces Tikhonov smooth function and apparent optical parameters to build an objective fitness function and weaken the ill condition of the particle size inversion equation. It also introduces multi-population genetic algorithm to solve the premature convergence of genetic algorithms. The results show that the relative error of the milk fat simulation solution with a nominal diameter is -3.52%, which meets the national standard of ±8% and better than the relative error of -5.01% of the standard genetic algorithm. Thus, the improved MPGA can reconstruct particle size distribution, with a good reliability and stability.

Spectral-Spatial Genetic Algorithm-Based Unsupervised Band Selection for Hyperspectral Image Classification

Band selection (BS) can mitigate the “curse of dimensionality” problem and improve the performance of hyperspectral image (HSI) classification. Genetic algorithms (GAs) have been applied to the task of hyperspectral BS showing significant advantages compared with other literature methods. However, the traditional GAs-based methods often select sets of bands having residual redundancy due to the large search space related to hyperspectral BS and the limitation of premature convergence in GAs. Moreover, existing GAs-based methods often are supervised, and that needs a large number of labeled samples to compute the fitness value for assessing the quality of selected bands. In this article, an unsupervised BS approach based on an improved GA is proposed. A fitness function based on the fisher score combined with superpixel is designed for evaluating the discriminability of band subsets considering both spectral and spatial information. Then, modified genetic operations are constructed to restrain the search space and reduce the redundancy of selected bands. The performance of the proposed spectral-spatial GA-based BS method is evaluated on three HSIs. The experimental results demonstrate that the proposed method is superior to the traditional GA-based method and seven state-of-the-art unsupervised methods.

Escaping Local Optima Using Crossover With Emergent Diversity

Population diversity is essential for avoiding premature convergence in genetic algorithms (GAs) and for the effective use of crossover. Yet the dynamics of how diversity emerges in populations are not well understood. We use rigorous runtime analysis to gain insight into population dynamics and GA performance for the ( ${\mu +1}$ ) GA and the Jump test function. We show that the interplay of crossover followed by mutation may serve as a catalyst leading to a sudden burst of diversity. This leads to significant improvements of the expected optimization time compared to mutation-only algorithms like the (1 + 1) evolutionary algorithm. Moreover, increasing the mutation rate by an arbitrarily small constant factor can facilitate the generation of diversity, leading to even larger speedups. Experiments were conducted to complement our theoretical findings and further highlight the benefits of crossover on the function class.

RETRACTED ARTICLE: A self-adaptive genetic algorithm with improved mutation mode based on measurement of population diversity

Genetic algorithm (GA) is an important and effective method to solve the optimization problem, which has been widely used in most practical applications. However, the premature convergence of GA has unexpected effect on the algorithm’s performance, the main reason is that the evolution of outstanding individuals multiply rapidly will lead to premature loss of population’s diversity. To solve the above problem, a method to qualify the population diversity and similarity between adjacent generations is proposed. Then, according to the evaluation of population diversity and the fitness of individual, the adaptive adjustment of crossover and mutation probability is realized. The results of several benchmark functions show that the proposed algorithm can search the optimal solution of almost all benchmark functions and effectively maintain the diversity of the population. Compared with the existing algorithms, it has greatly improved the convergence speed and the global optimal solution.

Tissue-like P system based DNA-GA for clustering

In recent years, DNA GA algorithm is drawing attention from scholars. The algorithm combines the DNA encoding and Genetic Algorithm, which solve the premature convergence of genetic algorithms, the weak local search capability and binary Hamming cliff problems effectively.How to design a more effective way to improve the performance of DNA-GA algorithm is more worth studying. As is known to all,the tissue-like P system can search for the optimal clustering partition with the help of its parallel computing advantage effectivel. This paper is under this premise and presents DNA-GA algorithm based on tissue-like P systems (TPDNA-GA) with a loop structure of cells, which aims to combine the parallelism and the evolutionary rules of tissue-like P systems to improve performance of the DNA-GA algorithm. The objective of this paper is to use the TPDNA-GA algorithm to support clustering in order to find the best clustering center.This algorithm is of particular interest to when dealing with large and heterogeneous data sets and when being faced with an unknown number of clusters. Experimental results show that the proposed TPDNA-GA algorithm for clustering is superior or competitive to classical k-means algorithm and several evolutionary clustering algorithms.

Tissue-like P system based DNA-GA for clustering

In recent years, DNA GA algorithm is drawing attention from scholars. The algorithm combines the DNA encoding and Genetic Algorithm, which solve the premature convergence of genetic algorithms, the weak local search capability and binary Hamming cliff problems effectively.How to design a more effective way to improve the performance of DNA-GA algorithm is more worth studying. As is known to all,the tissue-like P system can search for the optimal clustering partition with the help of its parallel computing advantage effectivel. This paper is under this premise and presents DNA-GA algorithm based on tissue-like P systems (TPDNA-GA) with a loop structure of cells, which aims to combine the parallelism and the evolutionary rules of tissue-like P systems to improve performance of the DNA-GA algorithm. The objective of this paper is to use the TPDNA-GA algorithm to support clustering in order to find the best clustering center.This algorithm is of particular interest to when dealing with large and heterogeneous data sets and when being faced with an unknown number of clusters. Experimental results show that the proposed TPDNA-GA algorithm for clustering is superior or competitive to classical k-means algorithm and several evolutionary clustering algorithms. Full Text: PDF DOI: http://dx.doi.org/10.11591/telkomnika.v16i3.9141

Optimal Configuration of PMU in Power System

In order to use the least number of phasor measurement unit (PMU) to guarantee the power system completely observable, an optimal PMU configuration method in power system was put forward. The pre-configuration of PMU was done considering the actual situation of power grid. The genetic algorithm (GA) was used for PMU configuration. Modify the formulae of crossover probability and mutation probability in traditional genetic algorithm to overcome the evolutionary stagnation when the maximum fitness value and the average fitness value in group were equal. The improved adaptive genetic algorithm (IAGA) was obtained. In order to eliminate the premature convergence of GA resulted from the chance and randomness of the crossover operation and mutation operation, the preventing premature operation was introduced. This method combined the IAGA and the preventing premature operation. It has good global astringency, and it can ensure the network complete observability with the minimum number of PMU.

Network Coding Optimization Method Research Based on Genetic Algorithm

Network coding optimization method research based on genetic algorithm applies network coding technology in monophyletic multicast network. After reaching network multicast rate, find link coding scheme which makes the minimum of the total number of network coding. Moreover, it makes the analysis and improvement for general genetic algorithm’s defects in network coding link optimization such as rare individual successful decoded by randomly generated initial population strategies, reduced algorithm search ability, premature convergence of genetic algorithm and long algorithm running time.

Generation bidding strategy in a pool based electricity market using Shuffled Frog Leaping Algorithm

In an electricity market generation companies need suitable bidding models to maximize their profits. Therefore, each supplier will bid strategically for choosing the bidding coefficients to counter the competitors bidding strategy. In this paper optimal bidding strategy problem is solved using a novel algorithm based on Shuffled Frog Leaping Algorithm (SFLA). It is memetic meta-heuristic that is designed to seek a global optimal solution by performing a heuristic search. It combines the benefits of the Genetic-based Memetic Algorithm (MA) and the social behavior-based Particle Swarm Optimization (PSO). Due to this it has better precise search which avoids premature convergence and selection of operators. Therefore, the proposed method overcomes the short comings of selection of operators and premature convergence of Genetic Algorithm (GA) and PSO method. Important merit of the proposed SFALA is that faster convergence. The proposed method is numerically verified through computer simulations on IEEE 30-bus system consist of 6 suppliers and practical 75-bus Indian system consist of 15 suppliers. The result shows that SFLA takes less computational time and producing higher profits compared to Fuzzy Adaptive PSO (FAPSO), PSO and GA.

A hybrid of adaptive neuro-fuzzy inference system and genetic algorithm

Premature convergence is an important problem in evolutionary algorithms, in particular genetic algorithm. The diversity of the population is a very influence paprameter on premature convergence in genetic algorithm. In this paper, we attempt to improve the performance of genetic algorithms by providing a bi-linear allocation lifetime approach to label the chromosomes based on their fitness values. These labales applied within a set of fuzzy rules and adaptive neuro-fuzzy inference system genetic algorithm to select suitable sexual chromosomes for recombination. We have evaluated the proposed technique on several numerical functions by comparing its performance to the basic genetic algorithm. The results of our initial experiments demonstrate a clear advantage of the adaptive neuro-fuzzy inference system genetic algorithm over the other techniques.

Hybrid system for handling premature convergence in GA – Case of grammar induction

Grammar induction (also known as grammar inference or language learning) is the process of learning grammar from training data. This paper presents hybrid genetic algorithm (GA) Tool for handling premature convergence for the case of grammar induction. The tool is developed in Java and the results are tested for the set of 20 languages. It is an effort towards introduction of a hybrid GA system where the collection of more than one type of reproduction methods is used for the grammar induction. Paper also includes the discussion on the various approaches used by author for handling the premature convergence problem in using GA.

Based on GA Mixed with Trust Region Method Solving Nonlinear Least Square Problems

Nonlinear least square is one of the unconstrained optimization problems. In order to solve the least square trust region sub-problem, a genetic algorithm (GA) of global convergence was applied, and the premature convergence of genetic algorithms was also overcome through optimizing the search range of GA with trust region method (TRM), and the convergence rate of genetic algorithm was increased by the randomness of the genetic search. Finally, an example of banana function was established to verify the GA, and the results show the practicability and precision of this algorithm.

Research on Grid Resources Schedule Based on an Adaptive Distribute Parallel Genetic Algorithm

In this paper an improved adaptive parallel genetic algorithm is proposed to solve problems of grid resources distribution and matching, comparing with the traditional genetic algorithms, a new adaptive selection operator is introduced, which can prevent the premature convergence of genetic algorithm efficiently. Besides, in this paper, the migration strategy of the parallel genetic algorithm can prevent the population trapped in the local extreme. And a pc-cluster containing eight computers is constructed to execute the coarse-grained parallel genetic algorithm and series genetic algorithm, and different scale resources and tasks are tested on the pc-cluster. Several examples are provided to be examined and the results illustrate that the proposed algorithm has higher global optimization capability, computational efficiency and stronger stability than the traditional genetic algorithm for the max time span. From these results, the parallel genetic algorithm reduced the searching time much more than series genetic algorithm for the same solutions. Moreover, compared with series genetic algorithm, the parallel genetic algorithm can get the more optimal solutions when the iteration is same.

Fast Evolution by Multiple Offspring Competition for Genetic Algorithms

The premature convergence of genetic algorithms (GAs) is the most major factor of slow evolution of GAs. In this paper we propose a novel method to solve this problem through competition of multiple offspring of in dividuals. Unlike existing methods, each parents in our method generates multiple offspring and then generated multiple offspring compete each other, finally winner offspring become to real offspring. From this multiple offspring competition, our GA rarel falls into the premature convergence and easily gets out of the local optimum areas without negative effects. This makes our GA fast evolve to the global optimum. Experimental results with four function optimization problems showed that our method was superior to the original GA and had similar performances to the best ones of queen-bee GA with best parameters.

A method of optimization for the distorted model of star map based on improved genetic algorithm

In order to realize the high precision attitude determination of the spacecrafts by star sensors, it is necessary to optimize accurately and rapidly the modeling of distorted star map. For the model, there are some shortcomings in traditional optimization methods, which result in large errors, slow speed and apt to falling into local optimal, etc. In this paper, an optimization method of star map distorted model based on improved genetic algorithm is presented. Different from the previous genetic algorithm, the coded-decimal notation is adopted to increase the execute-speed and real-time performance, the fitness evaluation function is improved to avoid premature convergence of genetic algorithms and the cross and mutation probability is selected appropriately to optimize model of distortion. Semi-physics simulation results show that compared with the traditional genetic algorithm, this method not only improves the speed but also greatly enhances the accuracy of the model.

A novel stochastic search method for polygonal approximation problem

A novel stochastic search method (NSSM) is proposed for the polygonal approximation problem. NSSM incorporates the ranking selection scheme, which is initially developed for solving the premature convergence of genetic algorithms (GAs), into the traditional split-and-merge technique. For avoiding getting trapped in a local optimum, NSSM randomly selects the splitting points and the merging points and determines the selection probability using the ranking selection scheme. Three groups of digital curves, including the synthesized benchmark curves and the real image curves, are used to test the performance of NSSM. The experimental results show the higher performance over the other methods including the GA-based methods and the local search methods.

The Study of FBG Sensor Network Based on Genetic and Simulated Annealing Algorithm

The genetic and simulated annealing algorithm (GASA) is used for optimizing the wavelength-division multiplexed (WDM) fiber Bragg gratings (FBGs) sensor network to detect wavelength detection accuracy of a few micro-strains even when the spectrums of FBGs are in partially or completely overlapped. The method could avoid some troubles of traditional methods, such as arduous trial-and-error procedure, premature convergence in genetic algorithm (GA) and search blindness in simulated annealing algorithm(SA). Experiments show that the application of the GASA in the FBG sensor network can adjust the optimization population and seek the global optimal result. Experimental results demonstrated the strain wavelength detection accuracy up to I μs has been achieved and a linear response are presented. Also the method improves the calculation efficiency greatly.

Optimal operation of water supply systems with tanks based on genetic algorithm

In view of the poor water supply system's network properties, the system's complicated network hydraulic equations were replaced by macroscopic nodal pressure model and the model of relationship between supply flow and water source head. By using pump-station pressure head and initial tank water levels as decision variables, the model of optimal allocation of water supply between pump-sources was developed. Genetic algorithm was introduced to deal with the model of optimal allocation of water supply. Methods for handling each constraint condition were put forward, and overcome the shortcoming such as premature convergence of genetic algorithm; a solving method was brought forward in which genetic algorithm was combined with simulated annealing technology and self-adaptive crossover and mutation probabilities were adopted. An application example showed the feasibility of this algorithm.

Hybrid Strategy for Parameter Estimation and PID Tuning

Parameter estimation and PID timing are two crucial issues in control engineering. Classical methods either require some prior information or depend on some rules, especially they are short of generality and their performances are not satisfied in many engineering fields. Although genetic algorithm and simulated annealing approaches have gained much attention and applications during the past decades, it may cause the premature convergence of genetic algorithm and prohibitive time-consumption required for simulated annealing if executing them alone. In this paper, reasonably combining the parallel structure of genetic algorithm with the controllable jumping property of simulated annealing, a class of effective and general hybrid optimization strategy is proposed for parameter estimation and PID tuning. The proposed strategy is easy to be understood and implemented, and only a little pre-needed information is required. Numerical simulation results demonstrate that the hybrid strategy is of effectiveness, robustness on initial states, and adaptability on models or plants, and comparisons show that the hybrid strategy can achieve performances greatly better than those of pure genetic algorithm and classical methods.