Variable Genetic Algorithm Research Articles

E-procurement optimization in supply chain: A dynamic approach using evolutionary algorithms

The increasing dynamism of global markets, coupled with the occurrence of unpredictable events, has introduced substantial challenges in formulating efficient supply chain strategies. The inherent dynamic nature of logistic networks necessitates a departure from traditional supply chain methodologies. This study proposes an advanced solution for dynamic e-procurement utilizing evolutionary algorithms (EAs). In conventional supply chains involving buyers and suppliers, a critical challenge is identifying cost-efficient suppliers capable of fulfilling consumer demands amidst fluctuating prices and quantities. Traditional optimization techniques often fail to perform effectively under these dynamic conditions. Moreover, detecting changes during the optimization process is an additional hurdle in dynamic optimization problems. Recent advancements have demonstrated the efficacy of EAs in solving a variety of real-world dynamic optimization issues.This research introduces a novel evolutionary algorithmic framework that integrates the Hybrid Multipopulational Reinitialization Strategy (HMRS), with a proposed hybrid change detection mechanism (named Smirnov-based Multi-sensor Detection Mechanism (SMDM)) to address the dynamic e-procurement problems. The proposed framework enhances the algorithm’s adaptability and responsiveness to real-time changes within the e-procurement environment. By effectively detecting and responding to these variations, the framework aims to optimize procurement processes, ensuring efficiency and robustness in managing fluctuating requirements and conditions inherent to dynamic e-procurement scenarios. The empirical analysis presented underscores the superiority of Differential Evolution (DE) variants over Genetic Algorithm (GA) variants within the procurement context. The detailed empirical study validates the effectiveness of the proposed dynamic approach in addressing the challenges associated with dynamic e-procurement. Considering real-world parameter fluctuations, the proposed approach demonstrates significant resilience, positioning it as a robust and efficient solution for optimizing the e-procurement process and adeptly managing the complexities of the supply chain environment.

Mitigation of total harmonic distortion and flicker emission in the presence of harmonic loads by optimal siting and sizing of wind turbines and energy storage systems

In distribution networks (DNs), there are significant technical challenges due to the high penetration of renewable energy resources (RERs), especially wind turbines (WTs) and the presence of non-linear electrical loads. One of the most important challenges is power quality. In such networks, to mitigate flicker emission produced by WTs and network voltage harmonics, it is necessary to solve the planning problems considering power quality objectives and standards. For this aim, this paper attempts to improve DN performance by optimal siting and sizing of WTs and energy storage systems (ESSs). In the planning problem, to take into account the power quality issue, flicker emission, total harmonic distortion (THD), and network voltage deviation (NVD) are included in the optimization problem. To effectively solve the planning problem, which is a non-linear and non-convex optimization problem, three variants of genetic algorithm (GA) are used. Simulation results indicate that optimal sizing and siting of WTs not only significantly improve the power quality issue but also improves the network performance.

A timely efficient and emissions-aware multiobjective truck-sharing integrated scheduling model in container terminals

In line with the development of green port terminals and the necessity of environmentally friendly efficient use of handling and transportation equipment, criteria such as emission reduction, energy minimization, and efficiency in a container terminal are prominent issues, which in recent years, have been the main concern of container terminal managers in developing terminal operating systems. Therefore, in this study, a bi-objective mixed integer mathematical programing model of integrated scheduling for ships’ loading and unloading operation sequencing in container terminals is proposed, which considers the minimization of containers’ flow time, trucks’ emission, and energy consumption as well as sharing trucks among quay cranes and decreasing their empty trips. In this model trucks’ technical specifications and polluting levels are input parameters for estimating emission and energy consumption of operation, which consequently leads to two uniform parallel machine scheduling models. To derive the exact solution of the model, the epsilon-constraint method is applied. Nevertheless, due to the problem’s NP-hardness attribute, two variants of non-dominated sorting genetic algorithm and multiobjective particle swarm optimization metaheuristic algorithms are proposed to identify the approximate Pareto front of solutions. In large-size problems, these algorithms outperform the epsilon-constraint method by finding acceptable results in a reasonable time. This model, which could be embedded in terminal operating software, not only yields a practical optimum operational sequence but also measures the amount of energy and emission from trucks during loading and unloading operations. The results, which utilized the data collected from Shahid Rajaee port in southern Iran, indicate that this practical model as compared to the current procedure in the terminal, results in a significant improvement in energy and emission reduction and terminal flow time.

An efficient energy saving scheme using reinforcement learning for 5G and beyond in H-CRAN

Maximizing the energy saving is one of the most important metrics in 5G and Beyond (B5G) cellular mobile networks. In order to satisfy the diverse requirements of 5G/B5G in dynamic environments, Reinforcement Learning (RL) is proven as a viable approach for solving resource management problems, especially for 5G energy resources. In this paper, we propose to apply the Q-Learning (QL) Reinforcement technique in the Heterogeneous Cloud 5G Radio Access Network (H-CRAN) architecture in order to optimize the energy efficiency in 5G/B5G networks. We compare its results with the Genetic Algorithm variant using Transformation (TGA) and Particle Swarm Optimization (PSO) under high and low traffic demands. The experimental results reveal the efficiency of RL compared to TGA and PSO techniques in terms of energy efficiency and system capacity.

Solution of OPF Using GA Variant With Inclusion of FACTS Devices

Optimal power flow (OPF) is a tool used for minimization through that secured and cost-effective power systems is obtained with the inclusion of FACTS devices for existing power systems that enhance the power transfer ability of the power system, which reduces congestion. The current work proposed a solution to the OPF issue in power systems using genetic algorithm (GA). The proposed technique is implemented on an IEEE 30 bus system by considering the minimization of fuel cost and L-index functions. The results at the end of this paper pre-sent the efficacy of GA algorithm to solve OPF issues in power systems and FACTS devices while comparing it to other algorithms presented previous re-search.

Wireless sensor network deployment optimization for a smart farming application: comparison of two Multi-Objective Evolutionary Algorithms

As it is one of the main means of insuring food security, agriculture has employed different technologies such as the Internet of Things and wireless sensor networks (WSN), to improve the quality and quantity of agricultural products, while preserving natural resources. But unfortunately, in agricultural plots where we have large surfaces of interest, the optimization of node deployment in a WSN remains among the major problems to be solved. In this work, we proposed a WSN node deployment optimization model for an agricultural application according to classical constraints of coverage, over-coverage, connectivity and nodes number, in addition to the nodes separating distance constraint which affects the quality of physical parameters models. We have applied two variants of Multi-Objective Genetic Algorithms, the Non Sorting Genetic Algorithm II (NS-GA II) and the Strength Pareto Evolutionary Algorithm II (SPEA II). As a result, for a 100 m2 plot, both algorithms ensured a communication rate of 100% while SPEA 2 presented a lower sensor number and over-coverage rates with a smaller separating distance, and the execution time of NSGA II was shorter with 11 s. Besides, both of them were greedy in terms of computation time with the increase in the size of the plots.

Intelligent and Resource-Conserving Service Function Chain (SFC) Embedding

Network Function Virtualization (NFV) opens us great opportunities for network processing with higher resource efficiency and flexibility. In this respect, there is an increasing need for intelligent orchestration mechanisms, such that NFV can exploit its potential and live up to its promise. Genetic algorithms have emerged as a promising alternative to the proliferation of heuristic and exact methods for the Service Function Chain (SFC) embedding problem. To this end, we design and evaluate a genetic algorithm (GA), which computes efficient embeddings with runtimes on par with approximate methods. We present a GA model as state-space search in order to clarify the design choices of a GA. Our proposed GA utilizes a heuristic for the generation of the initial population, with the aim of directing the search towards the solution. Given the sensitivity of GAs on their various parameters, we introduce a parameter adjustment framework for GA fine-tuning. A comparative evaluation among a range of GA variants with diverse features sheds light on the impact of these features on SFC embedding efficiency. The GA variant that stands out is further benchmarked against a baseline greedy algorithm and a state-of-the-art heuristic. Our evaluation results indicate that the GA yields notable gains in terms of request acceptance and resource efficiency.

A power quality‐based planning framework for flicker minimization of wind turbines in distribution network

AbstractRenewable energy penetration in distribution networks, especially wind turbines (WTs) and photovoltaics (PVs), leads to an increase in power quality disturbances. One of the most important power quality issues is flicker produced by WTs. Here, to mitigate the flicker produced by WTs, distribution network planning (DNP) problem is solved concerning flicker minimization. For this aim, a weighted objective function is defined in which in addition to power losses, flicker emission is also considered. In the planning problem, siting and sizing of WTs as well as siting of PVs are investigated to simultaneously minimize power losses and flicker emission. Owing to the difficulty of this optimization problem, a new variant of genetic algorithm (GA), named elitist‐based GA (EGA), is proposed whose crossover operator works based on the elite chromosome. This variant provides a promising way to efficiently use information from the best solution to generate new solutions. Simulation results show that optimal siting and sizing of WTs can considerably improve the network parameters in terms of power losses and flicker emission. Moreover, simulation results show the efficiency and effectiveness of EGA compared to the other studied techniques.

A supply chain network design for blood and its products using genetic algorithm: A case study of Turkey

<p style='text-indent:20px;'>In the paper, a novel mathematical model for blood supply chain network and a variant of Genetic Algorithm for the problem are proposed. The proposed mathematical model is developed using Mixed-integer nonlinear programming. The model is designed as a location, allocation, and routing problem that includes determining locations of local blood centers, assigning transfusion and blood donation centers to predetermined local blood centers, and routing between them. Eventually, a case study is conducted in the Eastern Black Sea Region of Turkish Red Crescent to illustrate the practicality and validity of the proposed model.</p>

ForestTaxator: A tool for detection and approximation of cross-sectional area of trees in a cloud of 3D points

In this paper we propose a novel software, named ForestTaxator, supporting terrestrial laser scanning data processing, which for dendrometric tree analysis can be divided into two main processes: tree detection in the point cloud and development of three-dimensional models of individual trees. The usage of genetic algorithms to solve the problem of tree detection in 3D point cloud and its cross-sectional area approximation with ellipse-based model is also presented. The detection and approximation algorithms are proposed and tested using various variants of genetic algorithms. The work proves that the genetic algorithms work very well: the obtained results are consistent with the reference data to a large extent, and the time of genetic calculations is very short. The attractiveness of the presented software is due to the fact that it provides all necessary functionalities used in the forest inventory field. The software is written in C# and runs on the .NET Core platform, which ensures its full portability between Windows, MacOS and Linux. It provides a number of interfaces thus ensuring a high level of modularity. The software and its code are made freely available.

Algorithms for restoring disaster-struck seaport operations considering interdependencies between infrastructure availability and repair team assignments

This paper presents new algorithms for restoring seaport operations after a disaster and develops a model considering interdependencies to select an efficient course of action. The model prioritises the infrastructure to be repaired, identifies the equipment required and the number of repair teams to be deployed. This paper develops a new dynamic programming model to assign multicrew repair teams and shows that the solution is exact. This paper then develops a new variant of the Hungarian Algorithm by embedding an exploitation-exploration strategy to obtain an approximate solution for large-sized assignment problems. Furthermore, this paper solves the restoration problem in totality by accounting for interdependencies between marine/land-side infrastructure/equipment and repair team assignments. This paper also develops a new variant of Genetic Algorithm based on a deletion-mutation technique and explores reducing the computation time involved in solving optimisation problems. This paper applies the principles laid out to restore Pantoloan seaport in Indonesia which was struck by a tsunami. The approximate solution obtained by the extended Hungarian Algorithm for small problems is quicker and matches with the exact solution obtained by the new dynamic programming. In case of large-sized problems, the extended Hungarian Algorithm has been found to arrive at a solution which allows reopening the seaport 48 % sooner than the other algorithms. The new variant of Genetic Algorithm outperforms the Genetic Algorithm with Local Search, needing only 40 % of the computation time and the solution found to be particularly stable too.

Comparative evaluation of the performance of improved genetic algorithms and differential evolution for in-core fuel management of a research reactor

This paper presents a comparative evaluation of the performance of Genetic Algorithm (GA) and Differential Evolution (DE) algorithm applied to in-core fuel management of the DNRR research reactor. Two GA variants corresponding to two selection operators, i.e., tournament (GA1) and roulette wheel (GA2) selections, respectively, with two-point crossover and scramble mutation were implemented for the ICFM problem. A comprehensive survey of the GA control parameters such as population size, crossover-type, mutation probability, and elitist archive size has been conducted to optimize the performance of the GAs. The basic DE was implemented with a standard mutation strategy DE/rand/1/bin. Numerical computations were performed based on the DNRR research reactor core loaded with 100 highly enriched uranium fuel (HEU) bundles for evaluating the performance of the GA and DE algorithms. Two main objectives were included in the fitness function to maximize the fuel cycle length and flatten the power distribution. The performance of the two GA variants and the basic DE was investigated with the same population size, fitness function, and convergence criterion. Each method was performed with 50 independent runs, and the best fitness values were collected for statistical analysis using Kruskal–Wallis and Mann–Whitney tests in comparison among the three methods. The statistical analysis shows that the performance of GA1 with tournament selection and DE are not significantly different and are better than GA2 with roulette wheel selection. DE is stable and efficient in exploring the search space to approach the global optimal solution in most runs. While, GA1 and GA2 were trapped at local optima by about 26% and 38%, respectively. However, the best solutions obtained with GA1 and GA2 after 50 independent runs are better than that obtained with DE in term of fitness values. This suggests an improvement of the basic DE is needed to maintain the potential good solutions during the evolution process.

A Variant Genetic Algorithm for a Specific Examination Timetabling Problem in a Japanese University

Background: Examination Timetabling Problem which tries to find an optimal examination schedule for schools, colleges, and universities, is a well-known NP-hard problem. This paper presents a Genetic Algorithm variant approach to solve a specific examination timetabling problem common in Japanese colleges and universities. Methods: The proposed algorithm uses direct chromosome representation Genetic Algorithm and implements constraint-based initialization and constraint-based crossover operations to satisfy the hard and soft constraints. An Island model with varying crossover and mutation probabilities and an improvement approach called pre-training are applied to the algorithm to further improve the result quality. Results: The proposed model is tested on synthetic as well as real datasets obtained from Sophia University, Japan and shows acceptable results. The algorithm was fine-tuned with different penalty points combinations and improvement combinations. Conclusion: The comparison results support the idea that the initial population pre-training and the island model are effective approaches to improve the result quality of the proposed model. Although the current island model used only four islands, incorporating greater number of islands, and some other diversity maintenance approaches such as memetic structures are expected to further improve the diversity and the result quality of the proposed algorithm on large scale problems.

GA-based geometrically optimized topology robustness to improve ambient intelligence for future internet of things

This study presents an ambient intelligence oriented, topological robustness scheme for internet of things (IoT). The scheme primarily exploits the underlying geometric properties of scale-free IoT networks for a substantial improvement of genetic algorithm (GA) based state of the art robustness techniques. The geometrically optimized GA (Go-GA) is subsequently extended to traditional heuristics algorithms by proposing their geometrically optimized variants. All three techniques are comparatively evaluated over a simulated scale-free IoT architecture employing Schneider R as metric of robustness. The study follows a data-driven approach where information about nodes and edges is pulled from a central big data server, and topological robustness of a given scale-free IoT is tested against existing benchmarks. The proposed scheme aims to achieve convergence to global optima and conserve computational costs by efficient edge swapping (EES) and node removal based thresholding (NRT). Performance evaluations show that Go-GA outperforms state of the art variants of GA by a margin of 20% for Schneider R. Traditional techniques of hill climbing algorithm (HCA), simulated annealing algorithm (SAA) and ROSE also improve by a margin of 11%, 12% and 14% respectively with consideration of geometric aspects. Moreover, as the network size increases, a mere decline of 7.6% in robustness R is observed for Go-GA as compared to 18% degradation for classical algorithms.

GPGPU based Dual Population Genetic Algorithm for solving Constrained Optimization Problem

Dual Population Genetic Algorithm is a variant of Genetic Algorithm that provides additional diversity to the main population. It covers the premature convergence problem as well as the diversity problem associated with Genetic Algorithm. But also its additional population introduces large search space that increases time required to find an optimal solution. This large scale search space problem can be easily solved using consumer-level graphics cards. The solution obtained using accelerated DPGA for solving a constrained optimization problem from CEC 2006 is compared with the obtained solution using sequential algorithm. The results show speed up maintaining solution quality.

Intelligent and Improved Self-Adaptive Anomaly based Intrusion Detection System for Networks

With the advent of digital technology, computer networks have developed rapidly at an unprecedented pace contributing tremendously to social and economic development. They have become the backbone for all critical sectors and all the top Multi-National companies. Unfortunately, security threats for computer networks have increased dramatically over the last decade being much brazen and bolder. Intrusions or attacks on computers and networks are activities or attempts to jeopardize main system security objectives, which called as confidentiality, integrity and availability. They lead mostly in great financial losses, massive sensitive data leaks, thereby decreasing efficiency and the quality of productivity of an organization. There is a great need for an effective Network Intrusion Detection System (NIDS), which are security tools designed to interpret the intrusion attempts in incoming network traffic, thereby achieving a solid line of protection against inside and outside intruders. In this work, we propose to optimize a very popular soft computing tool prevalently used for intrusion detection namely Back Propagation Neural Network (BPNN) using a novel machine learning framework called “ISAGASAA”, based on Improved Self-Adaptive Genetic Algorithm (ISAGA) and Simulated Annealing Algorithm (SAA). ISAGA is our variant of standard Genetic Algorithm (GA), which is developed based on GA improved through an Adaptive Mutation Algorithm (AMA) and optimization strategies. The optimization strategies carried out are Parallel Processing (PP) and Fitness Value Hashing (FVH) that reduce execution time, convergence time and save processing power. While, SAA was incorporated to ISAGA in order to optimize its heuristic search. Experimental results based on Kyoto University benchmark dataset version 2015 demonstrate that our optimized NIDS based BPNN called “ANID BPNN-ISAGASAA” outperforms several state-of-art approaches in terms of detection rate and false positive rate. Moreover, improvement of GA through FVH and PP saves processing power and execution time. Thus, our model is very much convenient for network anomaly detection.

Optimum Design of Laminated Composites for Minimum Thickness by a Variant of Genetic Algorithm

Optimum Design of Laminated Composites for Minimum Thickness by a Variant of Genetic Algorithm

Triploid genetic algorithm for convolutional neural network-based diagnosis of mild cognitive impairment.

The diagnosis of mild cognitive impairment (MCI), which is deemed a formative phase of dementia, may greatly assist clinicians in delaying its headway toward dementia. This article proposes a deep learning approach based on a triploid genetic algorithm, a proposed variant of genetic algorithms, for classifying MCI converts and non-converts using structural magnetic resonance imaging data. It also explores the effect of the choice of activation functions and that of the selection of hyper-parameters on the performance of the model. The proposed work is a step toward automated convolutional neural networks. The performance of the proposed method is measured in terms of accuracy and empirical studies exhibit the preeminence of our proposed method over the existing ones. The proposed model results in a maximum accuracy of 0.97961. Thus, it may contribute to the effective diagnosis of MCI and may prove important in clinical settings.

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.

Hybrid Knowledge Extraction Framework Using Modified Adaptive Genetic Algorithm and BPNN

Fault diagnosis based on the expert system (ES) is still a research topic of manufacturing in the Industry 4.0 because of the stronger interpretability. As the core component of the ES, fault diagnosis accuracy is positively correlated to the precise of the knowledge base. But it is difficult for users to understand the knowledge obtained from the original dataset utilizing the existing knowledge extraction method. Therefore, it is of great significance to extract easy-to-understand and exact rules from the NN framework. This paper proposes a hybrid extraction framework to perform the rule extraction for overcoming this drawback. First, an improved adaptive genetic algorithm (GA) using a logistic function, namely LAGA, is proposed to solve the traditional GA’s insufficient prediction performance issue. Compared with the other three mainstream adaptive GAs, the experiment results of optimizing six selected test functions by these GA variants show that the LAGA algorithm’s convergence accuracy and speed have been greatly improved, especially for high latitude functions. On this basis, a rule extraction method based on symbol rule and NN, namely the LAGA-BP framework, is discussed in this manuscript to classify the real-valued attributes. This framework obtains hidden knowledge (knowledge refinement process) by NN and further transforms the acquired hidden knowledge into more easy-to-understand rule knowledge (rule extraction process). The execution of the LAGA-BP framework could be separated into two phases. The first phase is to optimize a back propagation NN (BPNN) using the LAGA and refine prediction classification knowledge over the optimized BPNN. In the second phase, an attribute reduction algorithm using multi-layered NN (SD algorithm) based on two different superposed networks is used on this framework to reduce data-set attributes and then uses the K-means clustering algorithm to extract the if-then rule from the simplified attributes. the Wisconsin breast cancer data-set is used as a case study to reveal the correctness and robustness of the proposed LAGA-BP method. Consulting relevant medical personnel and referencing relevant data shows that the rules extracted using this method help verify the diagnosis results, thus verifying the proposed framework’s feasibility and practicality.