Simple Genetic Algorithm Research Articles

A multi-population state optimization algorithm for rail crack fault diagnosis

Rails usually operate in complex environments, which makes them prone to mechanical failures. In order to better diagnose crack faults, a multi-population state optimization algorithm (MPVHGA) is proposed in this paper, which is used to solve the problems of low efficiency, easy precocity, and easy convergence of local optimal solutions in traditional genetic algorithms. The fault signal detection results show that the MPVHGA has the advantages of fast convergence rate, high stability, no stagnation, and no limitation of the number of fixed iterations. The average iterations number of MPVHGA in 100 independent iterations is about one-fifth of the traditional single genetic algorithm (SGA for short) and about one-third of the population state optimization algorithm (VHGA for short), and the total convergence number of the MPVHGA converges to 55 and 10 more than the SGA and VHGA, respectively, and the accuracy of its fault diagnosis can reach 95.04%. On the basis of improving the performance of simple genetic algorithms, this paper provides a new detection method for rail crack fault diagnosis, which has important practical value in engineering.

On the computation of the automorphisms group of low density parity check codes using genetic algorithm

<p>The genetic algorithm (GA) is an adaptive metaheuristic search method based on the process of evolution and natural selection theory. It is an efficient algorithm used for solving the combinatorial optimization problems, e.g., travel salesman problem (TSP), linear ordering problem (LOP), and job-shop scheduling problem (JSP). The simple GA applied takes a long time to reach the optimal solution, the configuration of the GA parameters is vital for a successful GA search and convergence to optimal solutions, it includes population size, crossover operator, and mutation operator rates. Also, very recently, many research papers involved the GA in coding theory, In particular, in the decoding linear block codes case, which has heavily contributed to reducing the complexity, and guaranting the convergence of searching in fewer iterations. In this paper, an efficient method based on the genetic algorithm is proposed, and it is used for computing the Automorphisms groups of low density parity check (LDPC) codes, the results of the aforementioned method show a significant efficiency in finding an important set of Automorphisms set of LDPC codes.</p>

Matching Algorithm for Compact Ride-sharing in Rural Area using Genetic Algorithm

To solve the last mile problem in rural areas, we deal with compact ride-sharing model. For reducing the computational cost, we use a simple GA method and compared the performance of the parameters. To simplify the problem, our algorithm use a simple linear distance and minimize the total traveling distance. A data set used for GA was based on the Inami area in Nanto City, Toyama Prefecture, that is a real case study region. The performance comparison experiment of the algorithm by changing parameters was carried out. Experimental results show that the algorithm is likely to give correct answers for up to four vehicles. It was also shown that the combination of population size and elite proportion for GA could reduce computational costs while ensuring accuracy. In addition, by ride-sharing was actually carried out using the obtained experimental result, it was shown that it could be applied to the actual route even if the calculation was carried out in the linear distance.

Simple Deterministic Selection-Based Genetic Algorithm for Hyperparameter Tuning of Machine Learning Models

Hyperparameter tuning is a critical function necessary for the effective deployment of most machine learning (ML) algorithms. It is used to find the optimal hyperparameter settings of an ML algorithm in order to improve its overall output performance. To this effect, several optimization strategies have been studied for fine-tuning the hyperparameters of many ML algorithms, especially in the absence of model-specific information. However, because most ML training procedures need a significant amount of computational time and memory, it is frequently necessary to build an optimization technique that converges within a small number of fitness evaluations. As a result, a simple deterministic selection genetic algorithm (SDSGA) is proposed in this article. The SDSGA was realized by ensuring that both chromosomes and their accompanying fitness values in the original genetic algorithm are selected in an elitist-like way. We assessed the SDSGA over a variety of mathematical test functions. It was then used to optimize the hyperparameters of two well-known machine learning models, namely, the convolutional neural network (CNN) and the random forest (RF) algorithm, with application on the MNIST and UCI classification datasets. The SDSGA’s efficiency was compared to that of the Bayesian Optimization (BO) and three other popular metaheuristic optimization algorithms (MOAs), namely, the genetic algorithm (GA), particle swarm optimization (PSO) and biogeography-based optimization (BBO) algorithms. The results obtained reveal that the SDSGA performed better than the other MOAs in solving 11 of the 17 known benchmark functions considered in our study. While optimizing the hyperparameters of the two ML models, it performed marginally better in terms of accuracy than the other methods while taking less time to compute.

Forest Disaster Detection Method Based on Ensemble Spatial–Spectral Genetic Algorithm

Remote sensing images change detection is the key technology for monitoring forest windfall damages. Genetic algorithm (GA) is a branch of intelligent optimization techniques available to contribute to the surveys of windstorm and wildfire detection in forest areas. However, traditional genetic algorithms remain challenging due to several issues, such as complex calculation, poor noise immunity, and slow convergence. Analysis at the spatial level allows classifications to utilize contextual, and hierarchical information of image objects in addition to solely using spectral information. Additionally, ensemble learning presents a possibility for improving classification accuracy. Ensemble classifiers combined with spatial-based GA offers a promising method (E-nGA) for automating the process of monitoring forest loss. The research in this paper is presented in four parts. First, block-matching and 3D filtering is performed to suppress noises while enhancing valuable information. The difference image is then generated using the image difference method. Afterward, context-based saliency detection and Fuzzy c-means algorithm are conducted on the difference image to reduce the search space. Finally, the proposed E-nGA is executed to further classify the pixels and produce the final change map. Our first proposition is to design improved genetic operators in the GA, relying not only on pixel values but also on spatial information. Our second proposition is to consider an ensemble classification model based on multiple vegetation features for decision integration. Six frequently-used classification methods, as well as the simple genetic algorithm, are executed to demonstrate the effectiveness of the proposed framework in improving the robustness and detection accuracy.

A scalable species-based genetic algorithm for reinforcement learning problems

Abstract Reinforcement Learning (RL) methods often rely on gradient estimates to learn an optimal policy for control problems. These expensive computations result in long training times, a poor rate of convergence, and sample inefficiency when applied to real-world problems with a large state and action space. Evolutionary Computation (EC)-based techniques offer a gradient-free apparatus to train a deep neural network for RL problems. In this work, we leverage the benefits of EC and propose a novel variant of genetic algorithm called SP-GA which utilizes a species-inspired weight initialization strategy and trains a population of deep neural networks, each estimating the Q-function for the RL problem. Efficient encoding of a neural network that utilizes less memory is also proposed which provides an intuitive mechanism to apply Gaussian mutations and single-point crossover. The results on Atari 2600 games outline comparable performance with gradient-based algorithms like Deep Q-Network (DQN), Asynchronous Advantage Actor Critic (A3C), and gradient-free algorithms like Evolution Strategy (ES) and simple Genetic Algorithm (GA) while requiring far fewer hyperparameters to train. The algorithm also improved certain Key Performance Indicators (KPIs) when applied to a Remote Electrical Tilt (RET) optimization task in the telecommunication domain.

Effect of the initial population construction on the DBMEA algorithm searching for the optimal solution of the traveling salesman problem

There are many factors that affect the performance of the evolutionary and memetic algorithms. One of these factors is the proper selection of the initial population, as it represents a very important criterion contributing to the convergence speed. Selecting a conveniently preprocessed initial population definitely increases the convergence speed and thus accelerates the probability of steering the search towards better regions in the search space, hence, avoiding premature convergence towards a local optimum. In this paper, we propose a new method for generating the initial individual candidate solution called Circle Group Heuristic (CGH) for Discrete Bacterial Memetic Evolutionary Algorithm (DBMEA), which is built with aid of a simple Genetic Algorithm (GA). CGH has been tested for several benchmark reference data of the Travelling Salesman Problem (TSP). The practical results show that CGH gives better tours compared with other well-known heuristic tour construction methods.

Path planning method of smart mobile robot based on genetic algorithm

Robot path planning is an important topic in the field of robots. Unlike previous methods which consider the smoothness of paths in the process of genetic algorithm, this paper presents a new method of robot path planning that separates the genetic algorithm process from the path smoothing process. First, a simple genetic algorithm with variable length encoding is designed to generate a better polyline path, and then a new type of spiral with shape parameters is introduced to smooth it to smooth a larger corner. Throughout the path planning process, only obstacle coordinates are input to select parameters adaptively to generate the robot’s walking path. The simulation results show that separating the genetic algorithm process from the path smoothing process reduces the complexity of the genetic algorithm itself, so the designed smoothing operation not only improves the smoothness of the path, but also reduces the length of the path.

Applying hybrid genetic algorithm to multi-mode resource constrained multi-project scheduling problems

ABSTRACT Multi-mode resource-constrained multi-project scheduling problems (MMRCMPSP) are cases with a precedence relationship among activities, capacity constraints of different execution modes for activities, and multiple resources for multiple projects. In this study, hybrid genetic algorithm (HGA) and heuristic approach are developed to solve MMRCMPSP problems with the aim of minimizing makespan. The proposed HGA contains eight combinations of four typical priority rules (earliest due date, shortest process time, minimum slack, and maximum total work content) and two heuristic methods (serial and parallel) for MMRCMPSP. A total of 48 instances of related MMRCMPSP are considered from available resources and used as test beds for performance evaluation. Results demonstrate that the proposed HGA with parallel method and minimum slack priority rule outperforms a simple genetic algorithm and three activity-mode priority rule combinations from the recent literature. In addition, the superiority of HGA becomes increasingly significant when problem complexity increases.

A hybrid method for fast and robust topology optimization

To accelerate the convergence rate and obtain robust optimal results with clear profiles of structural topologies, this paper proposes a hybrid multi-population genetic algorithm (MPGA) and bi-directional evolutionary structural optimization (BESO) method for structural topology optimization. Each element in the design domain is treated as an individual and the elemental sensitivity is taken as the fitness function of one individual. Based on these treatments, MPGA operators, including crossover, mutation, migration and selection, are modified to adapt to compliance minimization problems. Additionally, some key parameters are controlled to guarantee a convergent solution and to solve the structural unconnectivity problem. A case is used to verify the effectiveness and efficiency of the proposed method. The numerical results show that the proposed method is efficient, and compared with the BESO method and combined simple genetic algorithm and BESO method, the proposed method provides a powerful ability in searching for better robust solutions and improving convergence speed.

Lower Bounds for Non-Elitist Evolutionary Algorithms via Negative Multiplicative Drift.

A decent number of lower bounds for non-elitist population-based evolutionary algorithms has been shown by now. Most of them are technically demanding due to the (hard to avoid) use of negative drift theorems-general results which translate an expected movement away from the target into a high hitting time. We propose a simple negative drift theorem for multiplicative drift scenarios and show that it can simplify existing analyses. We discuss in more detail Lehre's (2010) negative drift in populations method, one of the most general tools to prove lower bounds on the runtime of non-elitist mutation-based evolutionary algorithms for discrete search spaces. Together with other arguments, we obtain an alternative and simpler proof of this result, which also strengthens and simplifies this method. In particular, now only three of the five technical conditions of the previous result have to be verified. The lower bounds we obtain are explicit instead of only asymptotic. This allows us to compute concrete lower bounds for concrete algorithms, but also enables us to show that super-polynomial runtimes appear already when the reproduction rate is only a (1-ω(n-1/2)) factor below the threshold. For the special case of algorithms using standard bit mutation with a random mutation rate (called uniform mixing in the language of hyper-heuristics), we prove the result stated by Dang and Lehre (2016b) and extend it to mutation rates other than Θ(1/n), which includes the heavy-tailed mutation operator proposed by Doerr et al. (2017). We finally use our method and a novel domination argument to show an exponential lower bound for the runtime of the mutation-only simple genetic algorithm on OneMax for arbitrary population size.

A Multiple Sieve Approach Based on Artificial Intelligent Techniques and Correlation Power Analysis

Side-channel analysis achieves key recovery by analyzing physical signals generated during the operation of cryptographic devices. Power consumption is one kind of these signals and can be regarded as a multimedia form. In recent years, many artificial intelligence technologies have been combined with classical side-channel analysis methods to improve the efficiency and accuracy. A simple genetic algorithm was employed in Correlation Power Analysis (CPA) when apply to cryptographic algorithms implemented in parallel. However, premature convergence caused failure in recovering the whole key, especially when plenty of large S-boxes were employed in the target primitive, such as in the case of AES. In this article, we investigate the reason of premature convergence and propose a Multiple Sieve Method (MS-CPA), which overcomes this problem and reduces the number of traces required in correlation power analysis. Our method can be adjusted to combine with key enumeration algorithms and further improves the efficiency. Simulation experimental results depict that our method reduces the required number of traces by and , compared to classic CPA and the Simple-Genetic-Algorithm-based CPA (SGA-CPA), respectively, when the success rate is fixed to . Real experiments performed on SAKURA-G confirm that the number of traces required for recovering the correct key in our method is almost equal to the minimum number that makes the correlation coefficients of correct keys stand out from the wrong ones and is much less than the numbers of traces required in CPA and SGA-CPA. When combining with key enumeration algorithms, our method has better performance. For the traces number being 200 (noise standard deviation ), the attacks success rate of our method is , which is much higher than the classic CPA with key enumeration ( success rate). Moreover, we adjust our method to work on that DPA contest v1 dataset and achieve a better result (40.04 traces) than the winning proposal (42.42 traces).

Deep Innovation Protection: Confronting the Credit Assignment Problem in Training Heterogeneous Neural Architectures

Deep reinforcement learning approaches have shown impressive results in a variety of different domains, however, more complex heterogeneous architectures such as world models require the different neural components to be trained separately instead of end-to-end. While a simple genetic algorithm recently showed end-to-end training is possible, it failed to solve a more complex 3D task. This paper presents a method called Deep Innovation Protection (DIP) that addresses the credit assignment problem in training complex heterogenous neural network models end-to-end for such environments. The main idea behind the approach is to employ multiobjective optimization to temporally reduce the selection pressure on specific components in multi-component network, allowing other components to adapt. We investigate the emergent representations of these evolved networks, which learn to predict properties important for the survival of the agent, without the need for a specific forward-prediction loss.

A Framework for Implementing Metaheuristic Algorithms Using Intercellular Communication.

Metaheuristics (MH) are Artificial Intelligence procedures that frequently rely on evolution. MH approximate difficult problem solutions, but are computationally costly as they explore large solution spaces. This work pursues to lay the foundations of general mappings for implementing MH using Synthetic Biology constructs in cell colonies. Two advantages of this approach are: harnessing large scale parallelism capability of cell colonies and, using existing cell processes to implement basic dynamics defined in computational versions. We propose a framework that maps MH elements to synthetic circuits in growing cell colonies to replicate MH behavior in cell colonies. Cell-cell communication mechanisms such as quorum sensing (QS), bacterial conjugation, and environmental signals map to evolution operators in MH techniques to adapt to growing colonies. As a proof-of-concept, we implemented the workflow associated to the framework: automated MH simulation generators for the gro simulator and two classes of algorithms (Simple Genetic Algorithms and Simulated Annealing) encoded as synthetic circuits. Implementation tests show that synthetic counterparts mimicking MH are automatically produced, but also that cell colony parallelism speeds up the execution in terms of generations. Furthermore, we show an example of how our framework is extended by implementing a different computational model: The Cellular Automaton.

Artificial Intelligence: Excellent Key for Developing E&P Oil Industry in Iraq

A Multidisciplinary study for increasing oil recovery has been made in the present paper. This work has been adopted in the Upper Sandstone member/Zubair formation in South Rumaila Oil Field. The work was achieved by using optimization techniques for determining the optimal future reservoir performance regarding to infill drilling. Two different methods of Genetic Algorithm used to optimize the number and locations of infill wells. The first method is simple adaptive genetic algorithm and the second one is the breeder adaptive. The main parameters depended in this study is the cumulative oil production obtained from the output of reservoir simulation software. These two methods of GA depend on using Net Present Value (NPV) as economic analysis as objective function. The optimal number of infill wells is three wells which have maximum cumulative oil production and maximum value of NPV. The same results from two GA methods have been obtained. The locations of these optimal infill wells located in the crest of the oil field.

Optimal sizing and allocation of renewable based distribution generation with gravity energy storage considering stochastic nature using particle swarm optimization in radial distribution network

This paper presents an optimal sizing and allocation of a renewable energy resource (RES) based distribution generation (DG) units with gravity energy storage (GES) in the radial distribution network (DN). The optimization technique Constriction Coefficient Particle Swarm Optimization (CPSO) is utilized to reduce the total energy loss, which is subjected to equality and inequality constraints. Different DG parameters are considered and evaluated to reduce energy losses in electricity DN. To reduce search space and computational burden, a sensitivity analysis is performed to determine the candidate buses for the placement of DGs. The stochastic nature of RES (solar and wind), load, and storage unit has been handle using the probabilistic technique. The suitable penetration level is so adjusted as to restrict RES output on a certain fraction of the system load for stability consideration. The load flow analysis is performed using a backward-forward sweep algorithm embedded in the probability framework. The proposed approach has been examined on four different cases on DN consisting of 33 buses and it has been found that a notable reduction in losses with improved voltage profile is obtained by optimal sizing and placing DG units at an appropriate location. Results obtained using the CPSO technique has been validated by comparing it with the Simple Genetic Algorithm (SGA) technique. Further, the results obtained in case 3 using GES technology have compared with the battery storage system.

Modeling and optimization of batch production based on layout and cutting problems under uncertainty

This paper presents modeling and optimization of batch production based on layout, cutting and project scheduling problems by considering scenario planning. In order to solve the model, a novel genetic algorithm with an improvement procedure based on variable neighborhood search (VNS) is presented. Initially, the model is solved in small sizes using Lingo software and the combined (proposed) genetic algorithm; then the results are compared. Afterwards, the model is solved in large sizes by utilizing the proposed algorithm and simple genetic algorithm. The main findings of this paper show: (1) The suggested algorithm is valid and able to achieve optimal and near-optimal solutions. This conclusion was made after proving the validity of the proposed method by solving a case study by employing the classical method (employing Lingo 11). And when the results were compared with the ones obtained by the proposed algorithm, they were found to be the same in both cases. (2) The combined genetic algorithm is more effective in obtaining optimal boundaries and the solutions close to them in all cases compared to the classical (simple) genetic algorithm. In other words, the main finding of this paper is a combined genetic algorithm to optimize batch production modeling problems, which is more efficient than the methods provided in the literature.

Optimisation of friction stir welding parameters using combined Taguchi L9 and genetic algorithm

'Friction stir welding' (FSW) is an eco-friendly, energy-efficient solid-state welding procedure invented for high strength alloys as well as materials that are otherwise difficult for welding through the traditional fusion welding methods. This research paper elaborates on a hybrid Taguchi-genetic algorithm that optimises the FSW process parameters to yield favourable electrical conductivity of aluminium alloy AA1350. The optimisation parameters considered are travel speed, rotational speed and tool tilt angle. First, the Taguchi method is applied to reduce the number of design experiments and find the optimal set for quality parameters of the system. Subsequently, the genetic algorithm is employed to search for the optimum set of design parameters by using Taguchi solution as the initial population. The results obtained by the proposed method were compared to the outcomes of the conventional Taguchi L9 method and the simple genetic algorithm. The best performance was obtained by the proposed method in 51 iterations, whereas simple genetic algorithm used 182 iterations for the same.

Optimisation of friction stir welding parameters using combined Taguchi L9 and genetic algorithm

'Friction stir welding' (FSW) is an eco-friendly, energy-efficient solid-state welding procedure invented for high strength alloys as well as materials that are otherwise difficult for welding through the traditional fusion welding methods. This research paper elaborates on a hybrid Taguchi-genetic algorithm that optimises the FSW process parameters to yield favourable electrical conductivity of aluminium alloy AA1350. The optimisation parameters considered are travel speed, rotational speed and tool tilt angle. First, the Taguchi method is applied to reduce the number of design experiments and find the optimal set for quality parameters of the system. Subsequently, the genetic algorithm is employed to search for the optimum set of design parameters by using Taguchi solution as the initial population. The results obtained by the proposed method were compared to the outcomes of the conventional Taguchi L9 method and the simple genetic algorithm. The best performance was obtained by the proposed method in 51 iterations, whereas simple genetic algorithm used 182 iterations for the same.

Experimental Study of Hybrid Genetic Algorithms for the Maximum Scatter Travelling Salesman Problem

We consider the maximum scatter travelling salesman problem (MSTSP), a travelling salesman problem (TSP) variant. The problem aims to maximize the shortest edge in the tour that travels each city only once in the given network. It is a very complicated NP-hard problem, and hence, exact solutions are obtainable for small sizes only. For large sizes, heuristic algorithms must be applied, and genetic algorithms (GAs) are observed to be very successful in dealing with such problems. In our study, a simple GA (SGA) and four hybrid GAs (HGAs) are proposed for the MSTSP. The SGA starts with initial population produced by sequential sampling approach that is improved by 2-opt search, and then it is tried to improve gradually the population through a proportionate selection procedure, sequential constructive crossover, and adaptive mutation. A stopping condition of maximum generation is adopted. The hybrid genetic algorithms (HGAs) include a selected local search and perturbation procedure to the proposed SGA. Each HGA uses one of three local search procedures based on insertion, inversion and swap operators directly or randomly. Experimental study has been carried out among the proposed SGA and HGAs by solving some TSPLIB asymmetric and symmetric instances of various sizes. Our computational experience reveals that the suggested HGAs are very good. Finally, our best HGA is compared with a state-of-art algorithm by solving some TSPLIB symmetric instances of many sizes. Our computational experience reveals that our best HGA is better.