The evolutionary learning algorithm in processing of ecological data, case-study: poplar tree-ring analysis

This work investigates the benefits of using different distribution functions in the evolutionary learning algorithms with respect to Artificial Neural Networks’ (ANNs) generalization ability. We examine two modification of the recently proposed network weight-based evolutionary algorithm (NWEA), by mixing mutation strategies based on three distribution functions at the chromosome and the gene levels. The utilization of combined search strategies in the ANNs training implies that different step sizes determined by mixed distributions will direct the evolution towards good generalized ANNs.KeywordsArtificial Neural NetworksLearningEvolutionary Algorithms

This chapter presents an evolutionary neural fuzzy network, designed using the functional-link-based neural fuzzy network (FLNFN) and a new evolutionary learning algorithm. This new evolutionary learning algorithm is based on a hybrid of cooperative particle swarm optimization and cultural algorithm. It is thus called cultural cooperative particle swarm optimization (CCPSO). The proposed CCPSO method, which uses cooperative behavior among multiple swarms, can increase the global search capacity using the belief space. Cooperative behavior involves a collection of multiple swarms that interact by exchanging information to solve a problem. The belief space is the information repository in which the individuals can store their experiences such that other individuals can learn from them indirectly. The proposed FLNFN model uses functional link neural networks as the consequent part of the fuzzy rules. This chapter uses orthogonal polynomials and linearly independent functions in a functional expansion of the functional link neural networks. The FLNFN model can generate the consequent part of a nonlinear combination of input variables. Finally, the proposed functional-link-based neural fuzzy network with cultural cooperative particle swarm optimization (FLNFN-CCPSO) is adopted in several predictive applications. Experimental results have demonstrated that the proposed CCPSO method performs well in predicting the time series problems.

A model of wavelet neural network (WNN) using a new evolutionary learning algorithm is proposed in this paper. This new evolutionary learning algorithm is based on a hybrid of Particle Swarm Optimization (PSO) and gradient descent algorithm (GD), and is thus called HGDPSO. The Particle Swarm Optimizer has previously been used to train neural networks and generally met with success. The advantage of the PSO over many of the other optimization algorithms is its relative simplicity and quick convergence. But those particles collapse so quickly that it exits a potentially dangerous property: stagnation, which state would make it impossible to arrive at the global optimum, even a local optimum. HGDPSO was proposed for neural network training to avoid premature and eliminate stagnation in PSO. The effectiveness of the HGDPSO based WNN is demonstrated through the classification of the fault signals in rotating machinery. The simulated results show its feasibility and validity.

An intelligent methodology for power load forecasting was developed. In this forecasting system, wavelet neural network techniques were used in combination with a new evolutionary learning algorithm. The new evolutionary learning algorithm introduced the Tabu Search Algorithm and Genetic Mutation Operator into Artificial Fish Swarm Algorithm (AFSA) to construct a hybrid optimizing algorithm, and is thus called ASFA-TSGM. The hybrid algorithm can greatly improve the ability of searching the global excellent result and the convergence property and accuracy. The effectiveness of the ASFA-TSGM based WNN was demonstrated through the power load forecasting. The simulated results show its feasibility and validity.KeywordsWavelet neural networksTabu searchGenetic mutation operatorAFSAPower load forecasting

We present a tactile intelligent sensory substitution system (TIS/sup 3/) as a novel tactile communication system with dialog-based tuning. TIS/sup 3/ consists of a tactile encoder (TE) which maps desired objects or patterns (P) onto spatio-temporal stimulation patterns (P') as a parallel stream of stimulation time courses, by means of a tactile stimulator array (TS) at selected skin location to elicit tactile perceptions (P*). The human subject evaluates and compares a percept P* with a given object P as association goal. A Learning Module (LM) for dialog based TE-tuning transforms these evaluations into TE-change signals. In a first step, the application of a dialog-based TE-tuning was successfully tested using a TS version with fifteen stimulators on the lower arm for regaining the corresponding stimulation pattern P' for a given tactile reference percept P*/sub ref/. Optimization of P* was achieved in less than 100 iteration steps by means of a micro evolutionary learning algorithm (MEA).

Genetic-based evolutionary learning algorithms, such as genetic algorithms (GAs) and genetic programming (GP), have been applied to information retrieval (IR) since the 1980s. Recently, GP has been...

This paper examines the problem of intelligent parametric forecasting and trajectory optimization for unmanned aircraft systems (UAS) using evolutionary algorithms and machine learning methods. The relevance of the study stems from the multi-criteria and high complexity of UAS trajectory generation processes, as well as the need for accurate and timely assessment of its flight parameters. This is particularly important for ensuring the reliability, safety, and efficient performance of flight missions in UAS operating conditions, including scenarios related to the operation of critical infrastructure facilities. The objective of the study is to improve the accuracy of trajectory parameter diagnostics and the reliability of parametric forecasting of UAS trajectories under conditions of uncertainty and the multi-criteria nature of the problem. The paper proposes a hybrid approach incorporating a genetic algorithm (GA), a particle swarm algorithm (PSO), and an XGBoost machine learning model that provides adaptive assessment of the quality of the generated solutions. A computational software package has been implemented, including selection, recombination, mutation, and elite inheritance mechanisms, as well as a machine learning module for validating route trajectories and associated parameters. A computational experiment was conducted, which compared the effectiveness of GA and PSO under various operating scenarios. Testing was performed on industry-specific datasets with varying numbers of iterations. The computational experiment revealed the advantage of the genetic algorithm, namely, a 14–17% improvement in the quality of design solutions. The results of the study demonstrate high adaptability and practical applicability in modeling, parametric forecasting, and routing tasks, and also indicate the potential for integration with intelligent UAS navigation and monitoring systems. The article's materials are of practical interest to specialists in the field of UAS development and operation, as well as to researchers working on multi-criteria route planning, parametric forecasting, and improving the reliability of UAS operations.

<p>LiDAR (Light Detection and Ranging) systems such as ALS (Airborne Laser Scanning) are increasingly being used in studies that analyse the forest structure and for the characterisation of their ecosystem processes. The main reason is their ability to provide an accurate three-dimensional description of the canopy structure, compared to other existing methods, such as passive sensors or photogrammetry. In addition, the high positional accuracy of ALS and their capacity of penetrating the canopy through small gaps in the forest canopy allow the estimation of parameters such as aboveground biomass, vegetation height, or leaf area index, among others. In forestry applications, the acquisition of these parameters usually requires a pre-processing analysis of the point clouds, which includes ground point filtering, Digital Terrain Model (DTM) and Canopy Height Model (CHM) derivation, tree detection, and segmentation, among other processes.  In the last decades, point cloud processing has benefited by the development of dedicated software packages such as LAStools, FUSION, or Terrascan, focused on obtaining DTM/CHM and LiDAR-derived metrics. However, the recent development of more sophisticated software packages, such as LidR or Pycrown, allow implementing novel and state-of-the-art algorithms as well as specific user-created functions.</p><p>The wide variety of licensed and open-source software packages for ALS data processing, together with the increasing diversity of existing algorithms and methodologies, has provoked a multitude of comparative analysis of the most widely used algorithms in the scientific literature.  However, given the recent development of the field, a robust and exhaustive review of the current use of these software and the related algorithms is still missing. In this contribution, we present a synthesis review of 613 papers on the use of software packages and algorithms for ALS processing used between 2016 and 2020. The review focuses in forest environments with a complex structure where the difference in elevation, slope, and the existence of multiple vegetation strata usually requires more complex and specialised algorithms. Therefore, three specific steps of LiDAR processing workflow were considered: ground point filtering, DTM interpolation and crown detection and segmentation. The results showed that ground point filtering (84% of the studies) is the most common step in ALS processing, compared to DTM interpolation (71%) and tree segmentation (36%). For the DTM interpolation step, TIN construction was the most used method (13%) compared to other methods such as ordinary kriging (3%). Conventional software packages that employ algorithms based on progressive TIN densification or hieratical robust interpolation approaches were the most commonly used in ground point filtering for DTM generation. Meanwhile, other user developed advanced algorithms were used more frequently in canopy segmentation processing, especially in those articles using datasets with high point densities (165.93 p/m<sup>2</sup> on average), compared to datasets processed with more general software solutions as FUSION (13.81 p/m2).</p>

This book chapter explores the application of machine learning (ML) and deep learning (DL) algorithms for real-time data processing and decision-making within the context of smart cities. As urban environments evolve, the ability to harness vast amounts of data from diverse sourcessuch as sensors, IoT devices, and digital platformshas become essential for optimizing urban operations and improving quality of life. The chapter delves into the challenges and opportunities presented by these technologies, focusing on issues such as scalability, real-time performance, data quality, and the integration of heterogeneous systems. It also examines the critical role of model explainability, transparency, and ethical considerations in AI-driven decision-making processes. Moreover, the chapter discusses the importance of addressing uncertainty in smart city decision-making through advanced decision support systems, enabling cities to adapt dynamically to changing conditions. By investigating these themes, the chapter aims to provide comprehensive insights into the future of smart city systems powered by ML and DL, highlighting their potential to revolutionize urban management.

The efficacy of evolutionary or reinforcement learning algorithms for continuous control optimization can be enhanced by including an additional neural network dedicated to features extraction trained through self-supervision. In this paper we introduce a method that permits to continue the training of the features extracting network during the training of the control network. We demonstrate that the parallel training of the two networks is crucial in the case of agents that operate on the basis of egocentric observations and that the extraction of features provides an advantage also in problems that do not benefit from dimensionality reduction. Finally, we compare different feature extracting methods and we show that sequence-to-sequence learning outperforms the alternative methods considered in previous studies.

All applications in current trends need to use Machine Learning and Big Data in this huge data world.Machine learning is a type of artificial intelligence that allows software applications to become more accurate at predicting outcomes without being explicitly programmed to do so.Machine learning use historical data as input to predict new output values.Small Scale and even large developers cannot maintain Machine Learning Algorithms and other big data processing techniques all by themselves.To harvest, process and analyze the data we need a data processing engine.Our project is to provide them with a machine learning engine which runs on.a specified port on their servers which collects the data sets in real time.This which can be queried using a simple Query Language and analyze the database on the query.The purpose of our project is to create a Data Processing Engine for Big Data Analytics. Machine learning and Big Data plays a major role in the current huge data world. Our project is to maintain a dedicated machine learning and data processing engine, so that even small developers without the knowledge of machine learning and big data analysis can produce expected results based on Machine Learning to make their application smooth. All applications in current trends need to use Machine Learning and Big Data in this huge data world. Machine learning is a subset of artificial intelligence that allows software applications to become more accurate at predicting outcomes without being explicitly programmed to do so. Machine learning uses historical data as input to predict new output values. Small Scale and even large developers cannot maintain Machine Learning Algorithms and other big data processing techniques all by themselves. To harvest, process and analyze the data we need a data processing engine. This project is to provide them a data processing engine which runs on a specified port on their servers which collects the data sets in real time. This can be queried using a simple Query Language and analyze the data and produce respected results using ML based on the query

This review explores the use of machine learning (ML) techniques for detecting pests and diseases in crops, which is a significant challenge in agriculture, leading to substantial yield losses worldwide. This study focuses on the integration of ML models, particularly Convolutional Neural Networks (CNNs), which have shown promise in accurately identifying and classifying plant diseases from images. By analyzing studies published from 2019 to 2024, this work summarizes the common methodologies involving stages of data acquisition, preprocessing, segmentation, feature extraction, and prediction to develop robust ML models. The findings indicate that the incorporation of advanced image processing and ML algorithms significantly enhances disease detection capabilities, leading to the early and precise diagnosis of crop ailments. This can not only improve crop yield and quality but also reduce the dependency on chemical pesticides, contributing to more sustainable agricultural practices. Future research should focus on enhancing the robustness of these models to varying environmental conditions and expanding the datasets to include a wider variety of crops and diseases. CNN-based models, particularly specialized architectures like ResNet, are the most widely used in the studies reviewed, making up 42.36% of all models, with ResNet alone contributing 7.65%. This highlights ResNet’s appeal for tasks that demand deep architectures and sophisticated feature extraction. Additionally, SVM models account for 9.41% of the models examined. The prominence of both ResNet and MobileNet reflects a trend toward architectures with residual connections for deeper networks, alongside efficiency-focused designs like MobileNet, which are well-suited for mobile and edge applications.

When evolutionary algorithms are applied to problems with computationally intensive fitness functions a limited budget of evaluations is usually available. For these types of problems minimizing the number of function evaluations becomes paramount, which can be achieved by using smaller population sizes and limiting the number of generations per run. Unfortunately this leads to a limited sampling of the problem space, which means finding adequate solutions is less likely. Evolutionary algorithms (EA) can be augmented with machine learners (ML) to more effectively explore the problem space. However, a “well-tuned” evolutionary algorithm strikes a balance between its constituent operators. Failure to do so could mean implementations that prematurely converge to inferior solutions or to not converge at all. One aspect of such “tuning” is the use of a proper selection pressure. Introducing a machine learner into an EA/ML hybrid introduces a new form of “emergent” selection pressure for which practitioners may need to compensate. This research shows two implementations of EA/ML hybrids that filter out inferior offspring based on knowledge inferred from better individuals have different emergent selection pressure characteristics.

Evolutionary algorithms (EAs) have become the standards and paradigms for solving inverse problems. However, their two inherited operations, namely, the crossover and mutation operations, are complicated and difficult, both in theory and in numerical implementations. In this regard, increasing efforts have been devoted to EAs which are based on probabilistic models (EAPMs) to overcome the shortcomings of available EAs. The population-based incremental learning (PBIL) is an EAPM; moreover, it can bridge the gap between machine learning and the EAs, hence enjoying several merits compared with other EAs. However, lukewarm efforts have been devoted to PBILs, especially the real coded PBILs, in the study of inverse problems in electromagnetics. In this regard, a novel real coded PBIL is being proposed in this paper. In the proposed real coded PBIL, a probability matrix is proposed to randomly generate a population, and the updating formulas for this probability matrix using the so far searched best solution and the best solution of the current population are introduced to strike a balance between convergence performance and solution quality. The proposed real coded PBIL algorithm is numerically experimented on several case studies and promising results are reported in this paper.

Recent advances in manufacturing technology, such as cyber–physical systems, industrial Internet, AI (Artificial Intelligence), and machine learning have driven the evolution of manufacturing architectures into integrated networks of automation devices, services, and enterprises. One of the resulting challenges of this evolution is the increased need for interoperability at different levels of the manufacturing ecosystem. The scope ranges from shop–floor software, devices, and control systems to Internet-based cloud-platforms, providing various services on-demand. Successful implementation of interoperability in smart manufacturing would, thus, result in effective communication and error-prone data-exchange between machines, sensors, actuators, users, systems, and platforms. A significant challenge to this is the architecture and the platforms that are used by machines and software packages. A better understanding of the subject can be achieved by studying industry-specific communication protocols and their respective logical semantics. A review of research conducted in this area is provided in this article to gain perspective on the various dimensions and types of interoperability. This article provides a multi-faceted approach to the research area of interoperability by reviewing key concepts and existing research efforts in the domain, as well as by discussing challenges and solutions.