Multiple Fitness Functions Research Articles

Ensemble classifiers using multi-objective Genetic Programming for unbalanced data

Genetic Programming (GP) can be used to design effective classifiers due to its built-in feature selection and feature construction characteristics. Unbalanced data distributions affect the classification performance of GP classifiers. Some fitness functions have been proposed to solve the class imbalance problem of GP classifiers. However, with the evolution of GP, single-objective GP classifiers evaluated by a single fitness function have poor generalization ability. Moreover, using the best evolved GP classifier for decision-making can easily lead to the possibility of misclassification. In this paper, multi-objective GP is used to optimize multiple fitness functions including AUC approximation (Wmw), Distance (Dist), and Complexity to evolve ensemble classifiers, which jointly determines the class labels of unknown instances. Experiments on sixteen datasets show that our multi-objective GP can significantly improve classification performance compared with single-objective GP, and our proposed ensemble classifiers evolved by multi-objective GP can further improve the classification performance than the single best GP classifier. Comparisons with six GP-based and five traditional machine learning algorithms show that our proposed approaches can achieve significantly better classification performance on most cases.

VAWGAN-NPOA: Energy-Aware Routing for Innocuous Data Transmission in WSN

The shielding from and prediction of harmful activity in the wireless sensor network (WSN) are vital research problems. To solve these problems, an Energy-Aware Routing using a Variational auto-encoder Wasserstein generative adversarial network with a Nomadic People Optimizer (VAWGAN-NPOA-EAR-WSN) is proposed in this paper for secure data transmission in WSN. The proposed Variational auto-encoder Wasserstein generative adversarial network (VAWGAN) selects the Cluster Head (CH) based on multiple objective fitness functions such as delay, distance, energy, cluster density, and traffic rate. After that, energy-aware trust path selection is selected based on parameters such as trust, connectivity, and degree of amenity using the proposed Nomadic People Optimization Algorithm (NPOA). The data are transmitted from the cluster head to the base station and vice versa via a perfect trust path. The proposed VAWGAN-NPOA-EAR-WSN method is done in MATLAB and its efficacy is assessed with performance metrics, such as energy consumption, network lifetime, throughput, etc. compared with existing such as DNN-WSN, EESNN-WSN, and ANN-WSN. Finally, the proposed method provides a lower energy consumption of 2 mJ compared with existing methods.

A Multi-Objective Optimal Design Method for Gravity Compensators With Consideration of Minimizing Joint Reaction Forces

Abstract This paper presents a multi-objective optimal design method for gravity compensators with consideration of minimizing the joint reaction forces. High performance of the gravity compensation is achieved while the joint reaction forces are kept to a minimum. In this method, the ratio of the compensated torque to the uncompensated torque and the maximum value of the joint reaction forces are formulated as cost functions in the optimization problem, which is solved by adopting the Pareto front of multiple fitness functions with a genetic algorithm. This work takes a spring four-bar mechanism as a gravity compensator for a case study. The theoretical models of a gravity compensator and a robot manipulator show that the proposed multi-objective optimal design allows for the achievement of smaller joint reaction forces than the original single-objective optimal design, while their gravity compensation performances are relatively the same. Moreover, a prototype of a 0.2-kg gravity compensator realized from the proposed method was also built. An experimental study with this prototype showed that the measured motor torque was reduced by up to 93% within a range of 3π/4.

GLOD: The Local Greedy Expansion Method for Overlapping Community Detection in Dynamic Provenance Networks

Local overlapping community detection is a hot problem in the field of studying complex networks. It is the process of finding dense clusters based on local network information. This paper proposes a method called local greedy extended dynamic overlapping community detection (GLOD) to address the challenges of detecting high-quality overlapping communities in complex networks. The goal is to improve the accuracy of community detection by considering the dynamic nature of community boundaries and leveraging local network information. The GLOD method consists of several steps. First, a coupling seed is constructed by selecting nodes from blank communities (i.e., nodes not assigned to any community) and their similar neighboring nodes. This seed serves as the starting point for community detection. Next, the seed boundaries are extended by applying multiple community fitness functions. These fitness functions determine the likelihood of nodes belonging to a specific community based on various local network properties. By iteratively expanding the seed boundaries, communities with higher density and better internal structure are formed. Finally, the overlapping communities are merged using an improved version of the Jaccard coefficient, which is a measure of similarity between sets. This step ensures that overlapping nodes between communities are properly identified and accounted for in the final community structure. The proposed method is evaluated using real networks and three sets of LFR (Lancichinetti–Fortunato–Radicchi) networks, which are synthetic benchmark networks widely used in community detection research. The experimental results demonstrate that GLOD outperforms existing algorithms and achieves a 2.1% improvement in the F-score, a community quality evaluation metric, compared to the LOCD framework. It outperforms the best existing LOCD algorithm on the real provenance network. In summary, the GLOD method aims to overcome the limitations of existing community detection algorithms by incorporating local network information, considering overlapping communities, and dynamically adjusting community boundaries. The experimental results suggest that GLOD is effective in improving the quality of community detection in complex networks.

Social Learning and the Exploration-Exploitation Tradeoff

Cultures around the world show varying levels of conservatism. While maintaining traditional ideas prevents wrong ones from being embraced, it also slows or prevents adaptation to new times. Without exploration there can be no improvement, but often this effort is wasted as it fails to produce better results, making it better to exploit the best known option. This tension is known as the exploration/exploitation issue, and it occurs at the individual and group levels, whenever decisions are made. As such, it is has been investigated across many disciplines. We extend previous work by approximating a continuum of traits under local exploration, employing the method of adaptive dynamics, and studying multiple fitness functions. In this work, we ask how nature would solve the exploration/exploitation issue, by allowing natural selection to operate on an exploration parameter in a variety of contexts, thinking of exploration as mutation in a trait space with a varying fitness function. Specifically, we study how exploration rates evolve by applying adaptive dynamics to the replicator-mutator equation, under two types of fitness functions. For the first, payoffs are accrued from playing a two-player, two-action symmetric game, we consider representatives of all games in this class, including the Prisoner’s Dilemma, Hawk-Dove, and Stag Hunt games, finding exploration rates often evolve downwards, but can also undergo neutral selection as well depending on the games parameters or initial conditions. Second, we study time dependent fitness with a function having a single oscillating peak. By increasing the period, we see a jump in the optimal exploration rate, which then decreases towards zero as the frequency of environmental change increases. These results establish several possible evolutionary scenarios for exploration rates, providing insight into many applications, including why we can see such diversity in rates of cultural change.

Multi-View Point Cloud Registration Based on Evolutionary Multitasking With Bi-Channel Knowledge Sharing Mechanism

Multi-view point cloud registration is fundamental in 3D reconstruction. Since there are close connections between point clouds captured from different viewpoints, registration performance can be enhanced if these connections be harnessed properly. Therefore, this article models the registration problem as multi-task optimization, and proposes a novel bi-channel knowledge sharing mechanism for effective and efficient problem solving. The modeling of multi-view point cloud registration as multi-task optimization are twofold. By simultaneously considering the local accuracy of two point clouds as well as the global consistency posed by all the point clouds involved, a fitness function with an adaptive threshold is derived. Also a framework of the co-evolutionary search process is defined for the concurrent optimization of multiple fitness functions belonging to related tasks. To enhance solution quality and convergence speed, the proposed bi-channel knowledge sharing mechanism plays its role. The intra-task knowledge sharing introduces aiding tasks that are much simpler to solve, and useful information is shared across aiding tasks and the original tasks, accelerating the search process. The inter-task knowledge sharing explores commonalities buried among the original tasks, aiming to prevent tasks from getting stuck to local optima. Comprehensive experiments conducted on model object as well as scene point clouds show the efficacy of the proposed method.

Enhancing the network lifetime of a wireless sensor network using modified k-means Firefly Optimization

Developing technology including the Internet of Things, digital services, smart microgrids, and machine-to-machine systems encourage the implementation of self-configuring, automated systems in massive wireless sensor networks (WSNs). Effective Power utilization is critical in order to keep a network operational for the maximum possible time. Sensor nodes, which are small, battery-powered devices, are used in WSNs. As a result, one of the significant research issues in improving the lifespan of a WSN is resource management. A modified k-means (Mk-means) algorithm for clustering has been proposed to select Cluster Head (CH) to minimize energy usage of the nodes. Several optimization methods have been suggested in this area to extend the WSN lifetime. Following that, the Firefly Algorithm (FFA) is employed to generate the optimal routing through the CHs to a Base Station (BS), where multiple fitness functions such as residual energy, distance, and routing traffic are taken into account to optimize the FFA. As a result, information transmission among intermediate CHs in a hierarchy cluster-based design aids in lowering node energy usage. The proposed MKMFA (modified k-means Firefly Algorithm) technique’s performance is evaluated with K-means AODV (Ad-hoc On-demand Distance Vector) and IPC-KMAN(Improved Performance Clustering Using Modified K-Means Algorithm in Mobile Adhoc Networks) using network lifetime, throughput, Load Balancing Factor (LBF), and Packet Delivery Ratio (PDR). This work is based on recent advances in WSNs, which include application fields, design parameters, and lifetime prediction designs.

Optimizing the Energy Consumption of Blockchain-Based Systems Using Evolutionary Algorithms: A New Problem Formulation

Blockchain technology has gained recognition in industrial, financial, and various technological domains for its potential in decentralizing trust in peer-to-peer systems. A core component of blockchain technology is a consensus algorithm, most commonly Proof of Work (PoW). PoW is used in blockchain-based systems to establish trust among peers; however, it does require the expenditure of an enormous amount of energy that affects the environmental sustainability of blockchain-based systems. Energy minimization, whilst ensuring trust within blockchain-based systems that use PoW, is a challenging problem. The solution has to consider how energy consumption can be minimized without compromising trust, whilst still ensuring, for instance, scalability, security, and decentralization. In this paper, we represent the problem as a subset selection problem of miners in a blockchain-based system. We formulate the problem of blockchain energy consumption as a Search-Based Software Engineering problem with four objectives: energy consumption, carbon emission, decentralization, and trust. We propose a model composed of multiple fitness functions. The model can be used to explore the complex search space by selecting a subset of miners that minimizes the energy consumption without drastically impacting the primary goals of the blockchain technology (i.e., security/trustworthiness and decentralization). We integrate our proposed fitness functions into five evolutionary algorithms to solve the problem of blockchain miners selection. Our results show that the environmental sustainability of blockchain-based systems (e.g. reduced energy use) can be enhanced with little degradation in other competing objectives. We also report on the performance of the algorithms used.

Space optimizing design of ecological constructions based on attribute hierarchy model

Aiming at the problem of low user satisfaction of traditional ecological building space optimization methods, this paper introduces attribute analytic hierarchy process (AHP) model to effectively improve the effect of ecological building space optimization. This paper constructs an index system for the optimal design of ecological building space, and divides it into target level, criterion level and space level. Using the attribute analytic hierarchy process (AHP) model to construct attribute judgment matrix and weight, the general evaluation model of ecological building space optimization design is obtained. In this paper, multi-objective genetic algorithm is used to collect the spatial information of ecological building and determine the specific value of each parameter of ecological building. It used octopus to optimize the calculation results of multiple fitness functions and realized the optimized design of ecological building space. The experimental results show that the proposed method can significantly improve the user satisfaction of ecological building space, and people’s satisfaction of the three open spaces of the overhead fire fighting channel, the main building, and the arc wall enclosure space are 62.13 points, 43.93 points, and 52.55 points, respectively, which shows that the satisfaction of the sunken courtyard is the highest.

Genetic Algorithm Approach for Gains Selection of Induction Machine Extended Speed Observer

The subject of this paper is gains selection of an extended induction machine speed observer. A high number of gains makes manual gains selection difficult and due to nonlinear equations of the observer, well-known methods of gains selection for linear systems cannot be applied. A method based on genetic algorithms has been proposed instead. Such an approach requires multiple fitness function calls; therefore, using a quality index based on simulations makes gains selection a time-consuming process. To find a fitness function that evaluates, in a short time, quality indices based on poles placement have been proposed. As the observer is nonlinear, equations describing the observer dynamics have been linearized. The relationship between poles placement and real dynamic properties has been shown. A series of studies has been performed to investigate the influence of the operating point of the machine on the dynamics of the observer. It has been proven that rotor speed has a significant impact on the placement of the poles and the observer may lose stability after a rotation direction change. A method of gains modification to maintain symmetrical properties of the observer for both directions has been presented. Experimental studies of the observer during machine reverse in the open and closed-loop control system have been performed. The results show that the observer can be implemented in a sensorless drive, using the proposed gains selection method.

A Test Case Prioritization Genetic Algorithm Guided by the Hypervolume Indicator

Regression testing is performed during maintenance activities to assess whether the unchanged parts of a software behave as intended. To reduce its cost, test case prioritization techniques can be used to schedule the execution of the available test cases to increase their ability to reveal regression faults earlier. Optimal test ordering can be determined using various techniques, such as greedy algorithms and meta-heuristics, and optimizing multiple fitness functions, such as the average percentage of statement and branch coverage. These fitness functions condense the cumulative coverage scores achieved when incrementally running test cases in a given ordering using Area Under Curve (AUC) metrics. In this paper, we notice that AUC metrics represent a bi-dimensional (simplified) version of the hypervolume metric, which is widely used in many-objective optimization. Thus, we propose a Hypervolume-based Genetic Algorithm, namely HGA, to solve the Test Case Prioritization problem when using multiple test coverage criteria. An empirical study conducted with respect to five state-of-the-art techniques shows that (i) HGA is more cost-effective, (ii) HGA improves the efficiency of Test Case Prioritization, (iii) HGA has a stronger selective pressure when dealing with more than three criteria.

Rough set based semi-supervised feature selection via ensemble selector

Similar to feature selection over completely labeled data, the aim of feature selection over partially labeled data (semi-supervised feature selection) is also to find a feature subset which satisfies the intended constraint. Nevertheless, two difficulties may emerge in the semi-supervised feature selection: (1) labels are incomplete since labeled and unlabeled samples coexist in data; (2) the explanation of the selected feature subset is not clear. Therefore, such two problems will be mainly addressed in our research. Firstly, the unlabeled samples can be predicted through various semi-supervised learning methods. Secondly, the Local Neighborhood Decision Error Rate is proposed to construct multiple fitness functions for evaluating the significance of the candidate feature. Such mechanism not only realizes the ensemble selector in the process of feature selection, but also the qualified feature subset will bring us lower decision errors. Immediately, a heuristic algorithm is re-designed to execute feature selection. Finally, through testing nine different ratios (10%, 20%, …, 90%) of labeled samples in data, the experimental results demonstrate that our approach is superior to previous researches, mainly because: (1) the qualified feature subset derived by our approach can provide better classification performance; (2) the lower time consumption is required in our process of feature selection.

Evolutionary mining of skyline clusters of attributed graph data

Graph clustering is one of the most important research topics in graph mining and network analysis. Given the abundance of data in many real-world applications, graph nodes and edges could be annotated with multiple sets of attributes that could be derived from heterogeneous data sources. The consideration of these attributes during graph clustering would facilitate the generation of graph clusters with balanced and cohesive intra-cluster structures and nodes with homogeneous properties. In this paper, we propose a graph clustering approach for mining skyline clusters over large attributed graphs based on the dominance relationship. Each skyline solution is optimized simultaneously for multiple fitness functions, each function is defined over the graph topology or over a particular set of attributes derived from multiple data sources. We evaluate our approach experimentally with a large protein-protein interaction network of the human interactome enriched with large sets of heterogeneous cancer-associated attributes. The results demonstrate the efficiency of our approach and show how integrating node attributes from multiple data sources can result in a more robust graph clustering than the consideration of the graph topology alone.

Multi - focus image fusion using self - resemblance measure

Image fusion is the process of integrating several source images into a single image that provides more reliable information along with reduced redundancy. Inspired by the pixel correlation property, a fusion rule based on self-resemblance measure (SRM) is proposed. An efficient image fusion algorithm is presented for fusing multi-focus images using self resemblance measure, consistency verification and multiple objective particle swarm optimization. Initially based on SRM fusion rule, the source images are fused and then the block sizes are adaptively selected by particle swarm optimization based on the multiple objective fitness functions for obtaining enhanced fused image. The proposed framework is evaluated using quantitative metrics such as root mean square error, peak signal to noise ratio, correlation, standard deviation, mutual information and Petrovic metric. Experimental results demonstrate the outperformance of the proposed algorithm over many other well known state-of-the-art fusion techniques, both in visual effect and objective evaluation criteria.

A new approach for the design of digital frequency selective FIR filters using an FPA-based algorithm

Efficient digital filter design is an essential signal processing task. Finite Impulse Response (FIR) filters are used in many applications due to its properties of linear phase and frequency stability. Most traditional design methods suffer from the problem of insufficient control over the frequency response of the designed filter. For this reason, the use of a recently developed optimisation technique called flowers pollination algorithm (FPA), –based on the natural process of pollination of flowers– along with a novel multiple fitness function, is proposed in order to obtain optimised filter coefficients that best approximate ideal specifications. Results have been compared to both traditional methods (mainly windowing and the Parks-McClellan algorithm) as well as to several nature-inspired schemes. Finally, processing of a real EEG signal is used to quantitatively evaluate performance of designed filters. Numerical results show that our method achieves better fit to desired filter specifications, a 5−10 times larger attenuation in the stop band and a narrower transition band, at the expense of slightly increasing the pass-band ripple (5−15%) in 3 out of 4 of the cases.

Bi-stage hierarchical selection of pathway genes for cancer progression using a swarm based computational approach

BackgroundUnderstanding of molecular mechanism, lying beneath the carcinogenic expression, is very essential for early and accurate detection of the disease. It predicts various types of irregularities and results in effective drug selection for the treatment. Pathway information plays an important role in mapping of genotype information to phenotype parameters. It helps to find co-regulated gene groups whose collective expression is strongly associated with the cancer development. MethodIn this paper, we have proposed a bi-stage hierarchical swarm based gene selection technique which combines two methods, proposed in this paper for the first time. First one is a multi-fitness discrete particle swarm optimization (MFDPSO) based feature selection procedure, having multiple fitness functions. This technique uses multi-filtering based gene selection procedure. On top of it, a new blended Laplacian artificial bee colony algorithm (BLABC) is proposed and it is used for automatic clustering of the selected genes obtained from the first procedure. We have performed 10 times 10-fold cross validation and compared our proposed method with various statistical and swarm based gene selection techniques for different popular cancer datasets. ResultExperimental results show that the proposed method as a whole performs significantly well. The MFDPSO based system in combination with BLABC generates a good subset of pathway markers which provides more effective insight into the gene-disease association with high accuracy and reliability.

International Portfolio Diversification at Industry Level within South-East Asian Stock Markets

The issue of financial integration, at the country level, is a well-documented phenomenon in the area of International Portfolio Diversification (IPD). Despite the increasing degree of financial integration, it is important to investigate the global integration at industry level to capture the potential benefits of IPD. Thus, this study attempts to evaluate the potential advantages of IPD for international investors when investing in emergingstock markets of South-East Asia, through examining the co-integration within these markets at industry level during 2000-2012. Using Multiple Fitness Functions Genetic Algorithm (MFFGA) and co-integration techniques, the results imply that South-East Asian emergingstock markets are not co-integrated at the industry level, and thus great diversification gains can still be achieved by cross-industry portfolio diversification in this region. However, another contribution of the study is that the findings explicitly identify the industries that are better suited for diversification purposes.

A Space Division Multiobjective Evolutionary Algorithm Based on Adaptive Multiple Fitness Functions

The weighted sum of objective functions is one of the simplest fitness functions widely applied in evolutionary algorithms (EAs) for multiobjective programming. However, EAs with this fitness function cannot find uniformly distributed solutions on the entire Pareto front for nonconvex and complex multiobjective programming. In this paper, a novel EA based on adaptive multiple fitness functions and adaptive objective space division is proposed to overcome this shortcoming. The objective space is divided into multiple regions of about the same size by uniform design, and one fitness function is defined on each region by the weighted sum of objective functions to search for the nondominated solutions in this region. Once a region contains fewer nondominated solutions, it is divided into several sub-regions and one additional fitness function is defined on each sub-region. The search will be carried out simultaneously in these sub-regions, and it is hopeful to find more nondominated solutions in such a region. As a result, the nondominated solutions in each region are changed adaptively, and eventually are uniformly distributed on the entire Pareto front. Moreover, the complexity of the proposed algorithm is analyzed. The proposed algorithm is applied to solve 13 test problems and its performance is compared with that of 10 widely used algorithms. The results show that the proposed algorithm can effectively handle nonconvex and complex problems, generate widely spread and uniformly distributed solutions on the entire Pareto front, and outperform those compared algorithms.

Motor cortex neurons: from experiment to model via evolutionary algorithms

There are several types of neurons in primary motor cortex, with differences in ion channels, dynamics, morphology, as well as differences in synaptic input localization and axonal projections. To constrain computer models of cortical neurons, it is necessary to record from these different types of neurons. Model parameters can then be optimized, to minimize the difference between the model and experimental data. We have performed experiments recording from 2 classes of layer 5 pyramidal neurons in mouse motor cortex: corticospinal (SPI) and corticostriatal (STR). These neurons exhibit differences in synaptic inputs, axonal projections, and intrinsic membrane properties. Somatic whole-cell recordings were performed in both classes and current steps were injected (-0.2 to 0.6 nanoamperes) to produce different dynamical behaviors, including hyperpolarization-induced sag, alterations in firing rate and spike timing. We have developed multiple computer models of the different neuronal classes, each with varying complexity, and all constrained by experimental data. Two optimization methods were used: 1. evolutionary algorithms, 2. NEURON's principal axis (praxis) error minimization algorithm. We optimized model parameters with multiple fitness functions, first independently, and then in different combinations. The fitness functions quantified differences between experimental and simulated data, for firing frequency, resting membrane potential (RMP), action potential shape, interspike interval voltage trajectory, spike timing, hyperpolarization-induced sag, and voltage responses to onset/offset of current injections. We assessed optimized performance of different classes of neuronal compartmental models, with varying levels of morphological complexity, across the fitness functions. The models included a 5-compartment pyramidal neuron (with soma, basal dendrite, and three apical dendritic segments), and a ~904-compartment pyramidal neuron with detailed dendritic and axonal geometry reconstructed from our experiments. All model neurons included the same set of ion channels: INa and IKdr for action potential generation; IKa for rapid repolarization following an action potential; IKm and IKd for afterhyperpolarization; Ih for resonance, sag, excitability, and contribution to RMP; calcium channels (L, N, T-type) for bursting, excitability modulation, and contribution to backpropagating action potentials; calcium-activated potassium channels for regulating excitability after depolarization-induced calcium influx. We performed optimization in two phases: 1. optimize passive properties and compartmental geometry; 2. optimize maximal conductance and kinetics of each ion channel. The distribution of ion channels were constrained by patterns from experimental literature. For example, Ih increased exponentially with distance from the soma, while somatic and axonal compartments had higher densities of Na and K channels to allow action potential generation. This also reduced free parameters for optimization. Across the different models, the firing frequency, sag, and RMP had nearly perfect fits to the experimental data, while using a spike timing fitness function reduced spike time errors to within 1-10 ms. Due to the high dimensionality of parameter space, different subsets of fitness functions tended to improve independently, while others were correlated. We assessed which parameter values and ion channels had the largest impact on each fitness function using a post-optimization sensitivity analysis. Our methods were effective in optimizing both SPI and STR neurons and will likely generalize to other cell types.

Comparative Study of Genetic Algorithm Performed in a Single Generation for two Different Fitness Functions Technique f(x) = x^2 and f(x) = x^2+1

A genetic algorithm is one of a class of algorithms that searches a solution space for the optimal solution to a problem. First part of this work consists of basic information about Genetic algorithm like what are Individual, Population, Crossover, Genes, Binary Encoding, Flipping, Crossover probability, Mutation probability. What is it used for, what is their aim. In this article the methods of selection, crossover and mutation are specified. In the second part of this paper providing two different fitness functions f(x) = x and f(x) = x+1.Solving maximizing problems for two different fitness functions f(x) = x and f(x) = x+1 using genetic algorithm in a single generation. A single generation of a Genetic algorithm is performed here with encoding, selection, crossover and mutation. In this paper shown the best string from initial population is same (identical) for two different fitness functions f(x) = x and f(x) = x+1.The purpose of this paper is to present a specific varying fitness function (multiple fitness function) technique. The author of this paper was among the first that proposed the different fitness function technique used in GA for selecting the best string.