Combining a neural network with a genetic algorithm for process parameter optimization

A neural-network model has been developed to predict the value of a critical strength parameter (internal bond) in a particleboard manufacturing process, based on process operating parameters and conditions. A genetic algorithm was then applied to the trained neural network model to determine the process parameter values that would result in desired levels of the strength parameter for given operating conditions. The integrated NN–GA system was successful in determining the process parameter values needed under different conditions, and at various stages in the process, to provide the desired level of internal bond. The NN–GA tool allows a manufacturer to quickly determine the values of critical process parameters needed to achieve acceptable levels of board strength, based on current operating conditions and the stage of manufacturing.     

An optimization algorithm inspired by musical composition

In this paper we propose a new multiagent metaheuristic based in an artificial society that uses a dynamic creative system to compose music, called “Method of musical composition” or MMC. To show the performance of our proposed MMC algorithm, 13 benchmark continuous optimization problems and the related results are compared with harmony search, improved harmony search, global-best harmony search and self-adaptative harmony search. The experimental results demonstrate that MMC improves the results obtained by the other metaheuristics in a set of multi-modal functions.     

Animal migration optimization: an optimization algorithm inspired by animal migration behavior

In this paper, we intend to propose a new heuristic optimization method, called animal migration optimization algorithm. This algorithm is inspired by the animal migration behavior, which is a ubiquitous phenomenon that can be found in all major animal groups, such as birds, mammals, fish, reptiles, amphibians, insects, and crustaceans. In our algorithm, there are mainly two processes. In the first process, the algorithm simulates how the groups of animals move from the current position to the new position. During this process, each individual should obey three main rules. In the latter process, the algorithm simulates how some animals leave the group and some join the group during the migration. In order to verify the performance of our approach, 23 benchmark functions are employed. The proposed method has been compared with other well-known heuristic search methods. Experimental results indicate that the proposed algorithm performs better than or at least comparable with state-of-the-art approaches from literature when considering the quality of the solution obtained.     

Improved genetic algorithm inspired by biological evolution

The process of mutation has been studied extensively in the field of biology and it has been shown that it is one of the major factors that aid the process of evolution. Inspired by this a novel genetic algorithm (GA) is presented here. Various mutation operators such as small mutation, gene mutation and chromosome mutation have been applied in this genetic algorithm. In order to facilitate the implementation of the above-mentioned mutation operators a modified way of representing the variables has been presented. It resembles the way genetic information is coded in living beings. Different mutation operators pose a challenge as regards the determination of the optimal rate of mutation. This problem is overcome by using adaptive mutation operators. The main purpose behind this approach was to improve the efficiency of GAs and to find widely distributed Pareto-optimal solutions. This algorithm was tested on some benchmark test functions and compared with other GAs. It was observed that the introduction of these mutations do improve the genetic algorithms in terms of convergence and the quality of the solutions.     

Design optimization of a runflat structure based on multi-objective genetic algorithm

Runflat structure plays an important role in determining the sustainable mileage after the tire is shot. Lightweight, stiffness and strength are highly relevant to the overall performance of the structure. A parameterized model was built based on the full study of the structure, and a new adaptive meshing method is proposed to ensure the quality of the model. The accuracy of the new model was verified by comparing to the traditional finite element model. The parameter study was carried out to investigate the response of the performance and mass. Multi-objective optimization model was established by applying optimal Latin square design method and response surface model approach. Non-dominated sorting genetic algorithm-II (NSGA-II) was applied to obtain the optimization design. The results indicate that the combination of parameterized model and multi-objective genetic algorithms successfully achieve the goal of multi-objective optimization for mass and displacement while ensuring the stress. Meanwhile, the optimal topology, shape and thickness optimization for the runflat structure have been achieved at the same time.     

基于改进遗传算法的RBF神经网络结构优化研究

针对RBF神经网络隐含层节点数过多导致网络结构复杂的问题,提出了一种基于改进遗传算法(IGA)的RBF神经网络优化算法。利用IGA优化基于正交最小二乘法的RBF神经网络结构,通过对隐含层输出矩阵的列向量进行全局寻优,从而设计出结构更优的基于IGA的RBF神经网络(IGA-RBF)。将IGA-RBF神经网络的学习算法应用于电子元器件贮存环境温湿度预测模型,与基于正交最小二乘法的RBF神经网络进行比较的结果表明:IGA-RBF神经网络设计出来的网络训练步数减少了44步,隐含层节点数减少了34个,且预测模型得到的温湿度误差较小,拟合精度大于0.95,具有更高的预测精度。     

An optimization algorithm inspired by social creativity systems

The need for efficient and effective optimization problem solving methods arouses nowadays the design and development of new heuristic algorithms. This paper present ideas that leads to a novel multiagent metaheuristic technique based on creative social systems suported on music composition concepts. This technique, called “Musical Composition Method” (MMC), which was proposed in Mora-Gutiérrez et?al. (Artif Intell Rev 2012) as well as a variant, are presented in this study. The performance of MMC is evaluated and analyzed over forty instances drawn from twenty-two benchmark global optimization problems. The solutions obtained by the MMC algorithm were compared with those of various versions of particle swarm optimizer and harmony search on the same problem set. The experimental results demonstrate that MMC significantly improves the global performances of the other tested metaheuristics on this set of multimodal functions.     

Corona virus optimization (CVO): a novel optimization algorithm inspired from the Corona virus pandemic

This research introduces a new probabilistic and meta-heuristic optimization approach inspired by the Corona virus pandemic. Corona is an infection that originates from an unknown animal virus, which is of three known types and COVID-19 has been rapidly spreading since late 2019. Based on the SIR model, the virus can easily transmit from one person to several, causing an epidemic over time. Considering the characteristics and behavior of this virus, the current paper presents an optimization algorithm called Corona virus optimization (CVO) which is feasible, effective, and applicable. A set of benchmark functions evaluates the performance of this algorithm for discrete and continuous problems by comparing the results with those of other well-known optimization algorithms. The CVO algorithm aims to find suitable solutions to application problems by solving several continuous mathematical functions as well as three continuous and discrete applications. Experimental results denote that the proposed optimization method has a credible, reasonable, and acceptable performance.

     

ARO: A new model-free optimization algorithm inspired from asexual reproduction

This paper proposes a new individual based optimization algorithm, which is inspired from asexual reproduction known as a remarkable biological phenomenon, called as Asexual Reproduction Optimization (ARO). ARO can be essentially considered as an evolutionary based algorithm that mathematically models the budding mechanism of asexual reproduction. In ARO, each individual produces an offspring called bud through a reproduction mechanism; thereafter parent and its offspring compete according to a performance index obtained from the underlying objective function of the given optimization problem. This process leads to the fitter individual. ARO's adaptive search ability and its strong and weak points are described in this paper. Furthermore, the ARO convergence to the global optimum is mathematically analyzed. To approve the effectiveness of the ARO performance, it is tested with several benchmark functions frequently used in the area of optimization. Finally, the ARO performance is statistically compared with that of Particle Swarm Optimization (PSO). Results of simulation illustrate that ARO remarkably outperforms PSO.     

基于无约束优化和遗传算法的贝叶斯网络结构学习方法

基于无约束优化和遗传算法,提出一种学习贝叶斯网络结构的限制型遗传算法.首先构造一无约束优化问题,其最优解对应一个无向图.在无向图的基础上,产生遗传算法的初始种群,并使用遗传算法中的选择、交叉和变异算子学习得到最优贝叶斯网络结构.由于产生初始种群的空间是由一些最优贝叶斯网络结构的候选边构成,初始种群具有很好的性质.与直接使用遗传算法学习贝叶斯网络结构的效率相比,该方法的学习效率相对较高.     

改进遗传算法在磁性材料组炉优化问题中的应用

牌号、交货日期、优先级、需求量等是磁性材料生产工单的属性,计划员需要依据上述属性寻求最优的生产工单组合以最小化生产成本并提高生产效率.针对磁性材料企业人工组炉存在的组炉时间长,组炉结果不优化问题.本文建立了磁性材料生产工单组炉优化模型.提出将该组炉问题转化为伪旅行商问题,并采用一种改进遗传算法求解.染色体编码采用从1到N的自然数编码方式,并设计一种基于最早完工日期规则的初始种群产生方法.引入精英选择策略和改进的贪心三交叉算子,优化遗传算法收敛速度和精度;引入逆转算子,提高遗传算法全局搜索能力.基于实际生产数据的仿真实验表明,建立的磁性材料组炉优化模型是合适的,所提改进算法是有效的.     

A power spectrum optimization algorithm inspired by magnetotactic bacteria

Magnetotactic bacteria (MTB) are one kind of bacteria with magnetic particles called magnetosomes in their bodies. These particles often connect together like a chain. The MTB move toward the ideal living conditions under the interaction between magnetic field produced by the magnetic particles chain and that of the earth. In the paper, a new magnetic bacteria algorithm based on power spectrum (PSMBA) for optimization is proposed. The candidate solutions are decided by power spectrum in the algorithm. It mainly includes four steps: power spectrum calculation, bacteria swimming, bacteria rotation and bacteria replacement. The effect of swimming schemes and parameter settings on the performance of PSMBA is studied. And it is compared with GA, PSO and its variants and some other optimization algorithms on 25 benchmark functions including CEC2005. The simulation results show that PSMBA has better performance on most of the problems than most of the compared algorithms.     

A novel numerical optimization algorithm inspired from garden balsam

This paper introduces a new evolutionary computing method inspired by the seed transmission process of garden balsam. Garden balsam, a beautiful and attractive flower, randomly ejects the seeds within a certain range by virtue of mechanical force originating from cracking of mature seed pods, which is different from natural expansion of most species of plants. The seeds scattered to suitable growth area will have greater reproductive capacity in the next generation, followed by iteration until the most suitable point for growth in a particular space is eventually found. This phenomenon can more intuitively show the process of searching the problem solution space in the optimization problem. The garden balsam optimization algorithm proposed in this paper incorporates two different types of search processes and has a mechanism to maintain population diversity. Through the optimization experiment on 24 constrained optimization problems, the results obtained by using this algorithm are compared with those of some known meta-heuristic search algorithms. The statistical analysis of the experimental results has been implemented by Friedman rank test and Holm–Sidak test. The comparison results verify the effectiveness of the algorithm.

     

Intrusive tumor growth inspired optimization algorithm for data clustering

Inspired by the invasive tumor growth mechanism, this paper proposes a new meta-heuristic algorithm. A population of tumor cells can be divided into three subpopulations as proliferative cells, quiescent cells, and dying cells according to the nutrient concentration they get. Different cells have different behaviors and interactions among them for competition. In the tumor growing process, an invasive cell is born around a proliferative cell for the higher nutrient concentration and a necrotic cell occurs around a dying cell for the lower nutrient concentration, which presents the balance between life and death. To evaluate the performance of the intrusive tumor growth optimization algorithm (ITGO), we compared it to the many well-known heuristic algorithms by the Wilcoxon’s signed-rank test with Bonferroni–Holm correction method and the Friedman’s test. At the end, it is applied to solve the data clustering problem, which is a NP-hard problem. The experimental results show that the proposed ITGO algorithm outperforms other traditional heuristic algorithms for several benchmark datasets.     

Design of a parallel genetic algorithm for continuous and pattern-free heliostat field optimization

The Journal of Supercomputing - The heliostat field of solar power tower plants can suppose up to 50% of investment costs and 40% of energy loss. Unfortunately, obtaining an optimal field requires...     

Design optimization of discrete structural systems using MPI-enabled genetic algorithm

The focus of this paper is on the development and implementation of a genetic algorithm (GA)-based software system using message passing interface (MPI) protocol and library. A customized and improved form of simple GA used in previous research (Chen et al. 1997; Chen and Rajan 1998, 2000; Rajan et al. 1999) is parallelized. This MPI-enabled version is used to find the solution to finite element-based design optimization problems in a network of workstations. Results show that an almost linear speedup is obtained on homogenous hardware cluster and, with a proper load-balancing strategy, on heterogeneous hardware cluster.     

利用遗传模拟退火算法优化神经网络结构

常用的神经网络是通过固定的网络结构得到最优权值,使网络的实用性受到影响。引入了一种基于方向的交叉算子和变异算子,同时把模拟退火算法引入了遗传算法,结合遗传算法和模拟退火算法的优点,提出了一种优化神经网络结构的遗传——模拟退火混合算法,实现了网络结构和权值的同时优化。仿真实验表明,与遗传算法和模拟退火算法相比,该算法优化的神经网络收敛速度较快、预测精度较高,提高了网络的处理能力。     

递阶遗传粒子群算法在神经网络设计中的应用

将递阶遗传粒子群算法（HGAPSO）应用于神经网络设计,可以在对网络拓扑结构优化的同时对连接权重进行求解。该算法结合了遗传算法在解决离散问题和粒子群算法在解决连续问题上的优势,并利用BP算法沿误差最速下降的能力对连接权重进一步学习,达到全局最优和快速搜索的有机结合。通过对混沌时序信号的预测,表明递阶遗传粒子群算法在较大程度上提高了神经网络的学习性能和泛化能力。