A modified genetic algorithm for system optimization

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Alexander Zemliak
Keyword(s):  
Genetic Algorithm ◽  
Fitness Function ◽  
System Optimization ◽  
High Accuracy ◽  
Optimization Process ◽  
Processing Unit ◽  
Local Minima ◽  
Control Vector ◽  
Content Type ◽  
Cpu Time

Purpose In this paper, the previously developed idea of generalized optimization of circuits for deterministic methods has been extended to genetic algorithm (GA) to demonstrate new possibilities for solving an optimization problem that enhance accuracy and significantly reduce computing time. Design/methodology/approach The disadvantages of GAs are premature convergence to local minima and an increase in the computer operation time when setting a sufficiently high accuracy for obtaining the minimum. The idea of generalized optimization of circuits, previously developed for the methods of deterministic optimization, is built into the GA and allows one to implement various optimization strategies based on GA. The shape of the fitness function, as well as the length and structure of the chromosomes, is determined by a control vector artificially introduced within the framework of generalized optimization. This study found that changing the control vector that determines the method for calculating the fitness function makes it possible to bypass local minima and find the global minimum with high accuracy and a significant reduction in central processing unit (CPU) time. Findings The structure of the control vector is found, which makes it possible to reduce the CPU time by several orders of magnitude and increase the accuracy of the optimization process compared with the traditional approach for GAs. Originality/value It was demonstrated that incorporating the idea of generalized optimization into the body of a stochastic optimization method leads to qualitatively new properties of the optimization process, increasing the accuracy and minimizing the CPU time.

Analysis of strategies of circuit optimisation on basis of maximum principle

2018 ◽  
Vol 37 (1) ◽  
pp. 484-503
Author(s):  
Alexander Zemliak
Keyword(s):  
Maximum Principle ◽  
Dimensional Case ◽  
Processing Unit ◽  
Local Minima ◽  
Control Vector ◽  
Content Type ◽  
The Maximum Principle ◽  
Central Processing ◽  
Cpu Time ◽  
First Time

Purpose This paper aims to propose a new approach on the problem of circuit optimisation by using the generalised optimisation methodology presented earlier. This approach is focused on the application of the maximum principle of Pontryagin for searching the best structure of a control vector providing the minimum central processing unit (CPU) time. Design/methodology/approach The process of circuit optimisation is defined mathematically as a controllable dynamical system with a control vector that changes the internal structure of the equations of the optimisation procedure. In this case, a well-known maximum principle of Pontryagin is the best theoretical approach for finding of the optimum structure of control vector. A practical approach for the realisation of the maximum principle is based on the analysis of the behaviour of a Hamiltonian for various strategies of optimisation and provides the possibility to find the optimum points of switching for the control vector. Findings It is shown that in spite of the fact that the maximum principle is not a sufficient condition for obtaining the global minimum for the non-linear problem, the decision can be obtained in the form of local minima. These local minima provide rather a low value of the CPU time. Numerical results were obtained for both a two-dimensional case and an N-dimensional case. Originality/value The possibility of the use of the maximum principle of Pontryagin to a problem of circuit optimisation is analysed systematically for the first time. The important result is the theoretical justification of formerly discovered effect of acceleration of the process of circuit optimisation.

Analysis of the structure of different optimization strategies

2020 ◽  
Vol 39 (3) ◽  
pp. 583-593
Author(s):  
Alexander Zemliak ◽  
Jorge Espinosa-Garcia
Keyword(s):  
Lyapunov Function ◽  
Dynamic System ◽  
Optimal Structure ◽  
Optimization Process ◽  
Main Element ◽  
Circuit Optimization ◽  
Control Vector ◽  
Processor Time ◽  
Content Type ◽  
Cpu Time

Purpose In this paper, on the basis of a previously developed approach to circuit optimization, the main element of which is the control vector that changes the form of the basic equations, the structure of the control vector is determined, which minimizes CPU time. Design/methodology/approach The circuit optimization process is defined as a controlled dynamic system with a special control vector. This vector serves as the main tool for generalizing the problem of circuit optimization and produces a huge number of different optimization strategies. The task of finding the best optimization strategy that minimizes processor time can be formulated. There is a need to find the optimal structure of the control vector that minimizes processor time. A special function, which is a combination of the Lyapunov function of the optimization process and its time derivative, was proposed to predict the optimal structure of the control vector. The found optimal positions of the switching points of the control vector give a large gain in CPU time in comparison with the traditional approach. Findings The optimal positions of the switching points of the components of the control vector were calculated. They minimize processor time. Numerical results are obtained for various circuits. Originality/value The Lyapunov function, which is one of the main characteristics of any dynamic system, is used to determine the optimal structure of the control vector, which minimizes the time of the circuit optimization process.

scholarly journals FEM using projection of physical properties suitable for movement modeling and optimization processes

2020 ◽  
Vol 39 (5) ◽  
pp. 1185-1199
Author(s):  
Baptiste Ristagno ◽  
Dominique Giraud ◽  
Julien Fontchastagner ◽  
Denis Netter ◽  
Noureddine Takorabet ◽  
...  
Keyword(s):  
Three Dimensional ◽  
Transient State ◽  
Mathematical Function ◽  
Processing Unit ◽  
Time Saving ◽  
Content Type ◽  
Central Processing ◽  
Cpu Time ◽  
Simple Geometry ◽  
Evaluation Time

Purpose Optimization processes and movement modeling usually require a high number of simulations. The purpose of this paper is to reduce global central processing unit (CPU) time by decreasing each evaluation time. Design Methodology Approach Remeshing the geometry at each iteration is avoided in the proposed method. The idea consists in using a fixed mesh on which functions are projected to represent geometry and supply. Findings Results are very promising. CPU time is reduced for three dimensional problems by almost a factor two, keeping a low relative deviation from usual methods. CPU time saving is performed by avoiding meshing step and also by a better initialization of iterative resolution. Optimization, movement modeling and transient-state simulation are very efficient and give same results as usual finite element method. Research Limitations Implications The method is restricted to simple geometry owing to the difficulty of finding spatial mathematical function describing the geometry. Moreover, a compromise between imprecision, caused by the boundary evaluation, and time saving must be found. Originality Value The method can be applied to optimize rotating machines design. Moreover, movement modeling is performed by shifting functions corresponding to moving parts.

scholarly journals PENENTUAN PARAMETER LEARNING RATE SELAMA PEMBELAJARAN JARINGAN SYARAF TIRUAN BACKPROPAGATION MENGGUNAKAN ALGORITMA GENETIKA

2020 ◽  
Vol 14 (2) ◽  
pp. 202-212
Author(s):  
Ade chandra Saputra
Keyword(s):  
Neural Network ◽  
Genetic Algorithm ◽  
Genetic Algorithms ◽  
Fitness Function ◽  
Learning Rate ◽  
Local Minima ◽  
Rate Parameter ◽  
Neural Network Learning ◽  
Value Of Learning ◽  
Backpropagation Learning

One of the weakness in backpropagation Artificial neural network(ANN) is being stuck in local minima. Learning rate parameter is an important parameter in order to determine how fast the ANN Learning. This research is conducted to determine a method of finding the value of learning rate parameter using a genetic algorithm when neural network learning stops and the error value is not reached the stopping criteria or has not reached the convergence. Genetic algorithm is used to determine the value of learning rate used is based on the calculation of the fitness function with the input of the ANN weights, gradient error, and bias. The calculation of the fitness function will produce an error value of each learning rate which represents each candidate solutions or individual genetic algorithms. Each individual is determined by sum of squared error value. One with the smallest SSE is the best individual. The value of learning rate has chosen will be used to continue learning so that it can lower the value of the error or speed up the learning towards convergence. The final result of this study is to provide a new solution to resolve the problem in the backpropagation learning that often have problems in determining the learning parameters. These results indicate that the method of genetic algorithms can provide a solution for backpropagation learning in order to decrease the value of SSE when learning of ANN has been static in large error conditions, or stuck in local minima

Priority-based task scheduling on heterogeneous resources in the Expert Cloud

Kybernetes ◽  
2015 ◽  
Vol 44 (10) ◽  
pp. 1455-1471 ◽  
Author(s):  
Mehran Ashouraie ◽  
Nima Jafari Navimipour
Keyword(s):  
Genetic Algorithm ◽  
Task Scheduling ◽  
Design Methodology ◽  
Fitness Function ◽  
Content Type ◽  
Cloud Systems ◽  
New Class ◽  
Complete Problems ◽  
Np Complete ◽  
Real World Situation

Purpose – Expert Cloud as a new class of Cloud systems provides the knowledge and skills of human resources (HRs) as a service using Cloud concepts. Task scheduling in the Expert Cloud is a vital part that assigns tasks to suitable resources for execution. The purpose of this paper is to propose a method based on genetic algorithm to consider the priority of arriving tasks and the heterogeneity of HRs. Also, to simulate a real world situation, the authors consider the human-based features of resources like trust, reputation and etc. Design/methodology/approach – As it is NP-Complete to schedule tasks to obtain the minimum makespan and the success of genetic algorithm in optimization and NP-Complete problems, the authors used a genetic algorithm to schedule the tasks on HRs in the Expert Cloud. In this method, chromosome or candidate solutions are represented by a vector; fitness function is calculated based on several factors; one point cross-over and swap mutation are also used. Findings – The obtained results demonstrated the efficiency of the proposed algorithm in terms of time complexity, task fail rate and HRs utilization. Originality/value – In this paper the task scheduling issue in the Expert Cloud and improving pervious algorithm are pointed out and the approach to resolve the problem is applied into a practical example.

Ramjet Powered Missile Design Using a Genetic Algorithm

2007 ◽  
Vol 7 (2) ◽  
pp. 167-173 ◽  
Author(s):  
Roy J. Hartfield ◽  
Rhonald M. Jenkins ◽  
John E. Burkhalter
Keyword(s):  
Genetic Algorithm ◽  
Performance Prediction ◽  
Design Space ◽  
System Optimization ◽  
Design Strategy ◽  
Optimization Process ◽  
Preliminary Design ◽  
Efficient Exploration ◽  
Typical Design ◽  
Design Scenario

A methodology for developing optimized designs for symmetric-centerbody ramjet powered missiles, using a genetic algorithm as the driver for the system optimization process, has been developed. The methodology described in this paper allows for a comprehensive but efficient exploration of the design space. This global optimization process is made possible by performance prediction codes, which can provide preliminary design-level accuracy very efficiently. This work demonstrates the first truly comprehensive design strategy for this type of device. The paper contains a discussion of the methodology and shows results for a typical design scenario.

Acceleration of the numerical solution for the radiative transfer equation using a modified relaxation factor

2020 ◽  
Vol 37 (5) ◽  
pp. 1823-1847 ◽  
Author(s):  
Carlos Enrique Torres-Aguilar ◽  
Jesús Xamán ◽  
Pedro Moreno-Bernal ◽  
Iván Hernández-Pérez ◽  
Ivett Zavala-Guillén ◽  
...  
Keyword(s):  
Radiative Transfer ◽  
Numerical Solution ◽  
Transfer Equation ◽  
Radiative Transfer Equation ◽  
Boundary Surface ◽  
Processing Unit ◽  
Content Type ◽  
Cpu Time ◽  
Relaxation Factor ◽  
Factor Method

Purpose The purpose of this study is to propose a novel relaxation modified factor to accelerate the numerical solution of the radiative transfer equation (RTE) with several high-resolution total variation diminishing schemes. The methodology proposed is denoted as the X-factor method. Design/methodology/approach The X-factor method was compared with the technique deferred-correction (DC) for the calculations of a two-dimensional cavity with absorting-emiting-scatteting gray media using the discrete ordinates method. Four parameters were considered to evaluate: the absorption coefficient, the emissivity of boundary surface, the scattering albedo and under-relaxation factor. Findings The results showed the central processing unit (CPU) time of X-factor method was lower than DC. The reductions of CPU time with the X-factor method were observed from 0.6 to 75.4%. Originality/value The superiority of the X-factor method over DC was showed with the reduction of CPU time of the numerical solution of RTE for evaluated cases.

scholarly journals Real-time correction of antecedent precipitation for the Xinanjiang model using the genetic algorithm

2016 ◽  
Vol 18 (5) ◽  
pp. 803-815 ◽  
Author(s):  
Sheng-Li Liao ◽  
Gang Li ◽  
Qian-Ying Sun ◽  
Zhi-Fu Li
Keyword(s):  
Genetic Algorithm ◽  
Mean Squared Error ◽  
Fitness Function ◽  
Practical Method ◽  
Flood Forecasting ◽  
Optimization Process ◽  
Hunan Province ◽  
Runoff Simulation ◽  
Xinanjiang Model ◽  
Antecedent Precipitation

The Xinanjiang model has been successfully and widely applied in humid and semi-humid regions of China for rainfall–runoff simulation and flood forecasting. However, its forecasting precision is seriously affected by antecedent precipitation (Pa). Commonly applied methods relying on the experience of individual modelers are not standardized and difficult to transfer. In particular, the Xinanjiang daily model may result in obvious errors in the determination of Pa. Thus, a practical method for estimating Pa is proposed in this paper, which is based on a genetic algorithm (GA) and is estimated during a rising flood period. In the optimization process of a GA, Pa values form a chromosome, the root-mean-squared error between the observed and simulated streamflow is chosen as the fitness function. Simultaneously, the best individual reserved strategy is adopted between correction periods to avoid complete independence between each optimization process as well as to ensure the stability of the algorithm. Twenty-seven historical floods observed at the gauge station of the Shuangpai reservoir in Hunan Province of China are used to test the presented algorithm for estimation of Pa, and the results demonstrate that the proposed method significantly improves the quality of flood forecasting in the Xinanjiang model.

Job scheduling in the Expert Cloud based on genetic algorithms

Kybernetes ◽  
2014 ◽  
Vol 43 (8) ◽  
pp. 1262-1275 ◽  
Author(s):  
Nima Jafari Navimipour ◽  
Amir Masoud Rahmani ◽  
Ahmad Habibizad Navin ◽  
Mehdi Hosseinzadeh
Keyword(s):  
Genetic Algorithm ◽  
Cloud Computing ◽  
Job Scheduling ◽  
Fitness Function ◽  
Wait Time ◽  
Utilization Rate ◽  
Np Hard ◽  
Time Service ◽  
Content Type ◽  
Hard Problems

Purpose – Expert Cloud as a new class of Cloud computing systems enables its users to request the skill, knowledge and expertise of people by employing internet infrastructures and Cloud computing concepts without any information of their location. Job scheduling is one of the most important issue in Expert Cloud and impacts on its efficiency and customer satisfaction. The purpose of this paper is to propose an applicable method based on genetic algorithm for job scheduling in Expert Cloud. Design/methodology/approach – Because of the nature of the scheduling issue as a NP-Hard problem and the success of genetic algorithm in optimization and NP-Hard problems, the authors used a genetic algorithm to schedule the jobs on human resources in Expert Cloud. In this method, chromosome or candidate solutions are represented by a vector; fitness function is calculated based on response time; one point crossover and swap mutation are also used. Findings – The results indicate that the proposed method can schedule the received jobs in appropriate time with high accuracy in comparison to common methods (First Come First Served, Shortest Process Next and Highest Response Ratio Next). Also the proposed method has better performance in term of total execution time, service+wait time, failure rate and Human Resource utilization rate in comparison to common methods. Originality/value – In this paper the job scheduling issue in Expert Cloud is pointed out and the approach to resolve the problem is applied into a practical example.

An Automatic Registration Method for Multi-Modal Images Based on Alignment Metric

2012 ◽  
Vol 182-183 ◽  
pp. 1308-1312 ◽  
Author(s):  
Ying Jie Xia ◽  
Jian Qin Yin ◽  
Rui Chen
Keyword(s):  
Genetic Algorithm ◽  
Mutual Information ◽  
Fitness Function ◽  
High Accuracy ◽  
Automatic Registration ◽  
Registration Method ◽  
Iterative Search ◽  
Information Alignment

There are some methods such as mutual information, alignment metric and so on. But they aim at registration of multi-modal images whose sizes are same. And there are few researches on registration of multi-modal images whose sizes are different. In this paper, an automatic registration method for multi-modal images based on alignment metric is proposed. In order to look for the best registration location of two multi-modal images, genetic algorithm is used in the method for iterative search. Alignment metric is used as the fitness function of genetic algorithm. It is proved that automatic registration for multi-modal images of different sizes can be realized. And it has high accuracy and robustness.

Sign in / Sign up

Export Citation Format

Share Document