Varun Kumar Ojha

Metaheuristic design of feedforward neural networks

Over the past two decades, the feedforward neural network (FNN) optimization has been a key interest among the researchers and practitioners of multiple disciplines. The FNN optimization is often viewed from the various perspectives: the optimization of weights, network architecture, activation nodes, learning parameters, learning environment, etc. Researchers adopted such different viewpoints mainly to improve the FNNs generalization ability. The gradient-descent algorithm such as backpropagation has been widely applied to optimize the FNNs. Its success is evident from the FNNs application to numerous real-world problems. However, due to the limitations of the gradient-based optimization methods, the metaheuristic algorithms including the evolutionary algorithms, swarm intelligence, etc., are still being widely explored by the researchers aiming to obtain generalized FNN for a given problem. This article attempts to summarize a broad spectrum of FNN optimization methodologies including conventional and metaheuristic approaches. This article also tries to connect various research directions emerged out of the FNN optimization practices, such as evolving neural network (NN), cooperative coevolution NN, complex-valued NN, deep learning, extreme learning machine, quantum NN, etc. Additionally, it provides interesting research challenges for future research to cope-up with the present information processing era.

Design and Implementation of an Improved Datacenter Broker Policy to Improve the QoS of a Cloud

Cloud Computing offers various remotely accessible services to users either free or on payment. A major issue with Cloud Service Providers (CSP) is to maintain Quality of Service (QoS). The QoS encompasses different parameters, like, smart job allocation strategy, efficient load balancing, response time optimization, reduction in wastage of bandwidth, accountability of the overall system, best Virtual Machine (VM) (which reduce the overall execution time of the requested Cloudlets) selection etc. The Datacenter Broker (DCB) policy helps binding a Cloudlet with a VM. An efficient DCB policy reduces the overall execution time of a Cloudlet. Allocating cloudlets properly to the appropriate VMs in a Datacenter makes a system active, alive and balanced. In present study, we proposed a conductance algorithm for effective allocation of Cloudlets to the VMs in a Datacenter by taking into consideration of power and capacity of VMs, and length of Cloudlets. Experimental results obtained using CloudSim toolkit under heavy loads, establishes performance supremacy of our proposed algorithm over existing DCB algorithm.

Performance Analysis of Neuro Genetic Algorithm Applied on Detecting Proportion of Components in Manhole Gas Mixture

The article presents performance analysis of a real valued neuro genetic algorithm applied for the detection of proportion of the gases found in manhole gas mixture. The neural network (NN) trained using genetic algorithm (GA) leads to concept of neuro genetic algorithm, which is used for implementing an intelligent sensory system for the detection of component gases present in manhole gas mixture Usually a manhole gas mixture contains several toxic gases like Hydrogen Sulfide, Ammonia, Methane, Carbon Dioxide, Nitrogen Oxide, and Carbon Monoxide. A semiconductor based gas sensor array used for sensing manhole gas components is an integral part of the proposed intelligent system. It consists of many sensor elements, where each sensor element is responsible for sensing particular gas component. Multiple sensors of different gases used for detecting gas mixture of multiple gases, results in cross-sensitivity. The cross-sensitivity is a major issue and the problem is viewed as pattern recognition problem. The objective of this article is to present performance analysis of the real valued neuro genetic algorithm which is applied for multiple gas detection.

Ensemble of heterogeneous flexible neural trees using multiobjective genetic programming

Graphical abstractDisplay Omitted HighlightsA heterogeneous flexible neural tree (FNT) for function approximation was proposed.FNT was studied under Pareto-based multiobjective genetic programming framework.A diversity-index was introduced to maintain diversity in genetic population.FNT was found competitive with other algorithm when cross validated over datasets.Evolutionary weighted ensemble of HFNTs further improved FNT performance. Machine learning algorithms are inherently multiobjective in nature, where approximation error minimization and models complexity simplification are two conflicting objectives. We proposed a multiobjective genetic programming (MOGP) for creating a heterogeneous flexible neural tree (HFNT), tree-like flexible feedforward neural network model. The functional heterogeneity in neural tree nodes was introduced to capture a better insight of data during learning because each input in a dataset possess different features. MOGP guided an initial HFNT population towards Pareto-optimal solutions, where the final population was used for making an ensemble system. A diversity index measure along with approximation error and complexity was introduced to maintain diversity among the candidates in the population. Hence, the ensemble was created by using accurate, structurally simple, and diverse candidates from MOGP final population. Differential evolution algorithm was applied to fine-tune the underlying parameters of the selected candidates. A comprehensive test over classification, regression, and time-series datasets proved the efficiency of the proposed algorithm over other available prediction methods. Moreover, the heterogeneous creation of HFNT proved to be efficient in making ensemble system from the final population.

Predictive modeling of die filling of the pharmaceutical granules using the flexible neural tree

In this work, a computational intelligence (CI) technique named flexible neural tree (FNT) was developed to predict die filling performance of pharmaceutical granules and to identify significant die filling process variables. FNT resembles feedforward neural network, which creates a tree-like structure by using genetic programming. To improve accuracy, FNT parameters were optimized by using differential evolution algorithm. The performance of the FNT-based CI model was evaluated and compared with other CI techniques: multilayer perceptron, Gaussian process regression, and reduced error pruning tree. The accuracy of the CI model was evaluated experimentally using die filling as a case study. The die filling experiments were performed using a model shoe system and three different grades of microcrystalline cellulose (MCC) powders (MCC PH 101, MCC PH 102, and MCC DG). The feed powders were roll-compacted and milled into granules. The granules were then sieved into samples of various size classes. The mass of granules deposited into the die at different shoe speeds was measured. From these experiments, a dataset consisting true density, mean diameter (d50), granule size, and shoe speed as the inputs and the deposited mass as the output was generated. Cross-validation (CV) methods such as 10FCV and 5x2FCV were applied to develop and to validate the predictive models. It was found that the FNT-based CI model (for both CV methods) performed much better than other CI models. Additionally, it was observed that process variables such as the granule size and the shoe speed had a higher impact on the predictability than that of the powder property such as d50. Furthermore, validation of model prediction with experimental data showed that the die filling behavior of coarse granules could be better predicted than that of fine granules.

Understating continuous ant colony optimization for neural network training: A case study on intelligent sensing of manhole gas components

In this work, we proposed various strategies for improving the performance of continuous ant colony optimization algorithm (ACO∗), which was used here for optimizing neural network (NN). Here, a real-world problem, that is, detection of manhole gas components, was used for case study. Manhole contains various toxic and explosive gases. Therefore, pre-detection of these toxic gases is crucial to avoid human fatality. Hence, we proposed to design an intelligent sensory system, which used a trained NN for detecting manhole gases. The training to NN was provided using dataset that was generated using laboratory tests, sensor’s data-sheets, and literature. The primary focus of this work was on the performance evaluation and improvement of ACO∗ algorithm. Hence, understanding of ACO∗ parameter tuning and enhancements of ACO∗ parameters through performance evaluation was well studied. Moreover, complexity analysis of ACO∗ was firmly addressed. We extended our article scope to cover the performance comparisons between ACO∗ and other NN training algorithms. We found that the improved ACO∗ performed best in comparison to other NN training algorithms such as backpropagation, conjugate gradient, particle swarm optimization, simulated annealing, and genetic algorithm.

ACO for continuous function optimization: A performance analysis

The performance of the meta-heuristic algorithms often depends on their parameter settings. Appropriate tuning of the underlying parameters can drastically improve the performance of a meta-heuristic. The Ant Colony Optimization (ACO), a population based meta-heuristic algorithm inspired by the foraging behavior of the ants, is no different. Fundamentally, the ACO depends on the construction of new solutions, variable by variable basis using Gaussian sampling of the selected variables from an archive of solutions. A comprehensive performance analysis of the underlying parameters such as: selection strategy, distance measure metric and pheromone evaporation rate of the ACO suggests that the Roulette Wheel Selection strategy enhances the performance of the ACO due to its ability to provide non-uniformity and adequate diversity in the selection of a solution. On the other hand, the Squared Euclidean distance-measure metric offers better performance than other distance-measure metrics. It is observed from the analysis that the ACO is sensitive towards the evaporation rate. Experimental analysis between classical ACO and other meta-heuristic suggested that the performance of the well-tuned ACO surpasses its counterparts.

Ensemble of Heterogeneous Flexible Neural Tree for the Approximation and Feature-Selection of Poly (Lactic-co-glycolic Acid) Micro- and Nanoparticle

In this work, we used an adaptive feature-selection and function approximation model, called, flexible neural tree (FNT) for predicting Poly (lactic-co-glycolic acid) (PLGA) micro- and nanoparticle’s dissolution-rates that bears significant role in the pharmaceutical, medical, and drug manufacturing industries. Several factor influences PLGA nanoparticles dissolution-rate prediction. FNT model enable us to deal with feature-selection and prediction simultaneously. However, a single FNT model may or may not offer a generalized solution. Hence, to build a generalized model, we used an ensemble of FNTs. In this work, we have provided a comprehensive study for examining the most significant (influencing) features that influences dissolution rate prediction.

Dimensionality reduction, and function approximation of poly(lactic-co-glycolic acid) micro- and nanoparticle dissolution rate

Prediction of poly(lactic-co-glycolic acid) (PLGA) micro- and nanoparticles’ dissolution rates plays a significant role in pharmaceutical and medical industries. The prediction of PLGA dissolution rate is crucial for drug manufacturing. Therefore, a model that predicts the PLGA dissolution rate could be beneficial. PLGA dissolution is influenced by numerous factors (features), and counting the known features leads to a dataset with 300 features. This large number of features and high redundancy within the dataset makes the prediction task very difficult and inaccurate. In this study, dimensionality reduction techniques were applied in order to simplify the task and eliminate irrelevant and redundant features. A heterogeneous pool of several regression algorithms were independently tested and evaluated. In addition, several ensemble methods were tested in order to improve the accuracy of prediction. The empirical results revealed that the proposed evolutionary weighted ensemble method offered the lowest margin of error and significantly outperformed the individual algorithms and the other ensemble techniques.

Ensemble of Flexible Neural Trees for Predicting Risk in Grid Computing Environment

Risk assessment in grid computing is an important issue as grid is a shared environment with diverse resources spread across several administrative domains. Therefore, by assessing risk in grid computing, we can analyze possible risks for the growing consumption of computational resources of an organization and thus we can improve the organization’s computation effectiveness. In this paper, we used a function approximation tool, namely, flexible neural tree for risk prediction and risk (factors) identification. Flexible neural tree is a feed forward neural network model, where network architecture was evolved like a tree. Our comprehensive experiment finds score for each risk factor in grid computing together with a general tree-based model for predicting risk. We used an ensemble of prediction models to achieve generalization.