Feel free to browse the list of my
PUBLICATIONS

 

round_concept_cropped

2016

 

Title: An Online Structural Plasticity Rule for Generating Better Reservoirs

Authors: S. Roy and A. Basu

Published: Neural Computation, MIT Press

Abstract: In this article, a novel neuro-inspired low-resolution online unsupervised learning rule is proposed to train the reservoir or liquid of Liquid State Machine. The liquid is a sparsely interconnected huge recurrent network of spiking neurons. The proposed learning rule is inspired from structural plasticity and trains the liquid through formation and elimination of synaptic connections. Hence, the learning involves rewiring of the reservoir connections similar to structural plasticity observed in biological neural networks. The network connections can be stored as a connection matrix and updated in memory by using Address Event Representation (AER) protocols which are generally employed in neuromorphic systems. On investigating the 'pairwise separation property' we find that trained liquids provide 1.36 ± 0.18 times more inter-class separation while retaining similar intra-class separation as compared to random liquids. Moreover, analysis of the 'linear separation property' reveals that trained liquids are 2.05 ± 0.27 times better than random liquids. Furthermore, we show that our liquids are able to retain the 'generalization' ability and 'generality' of random liquids. A memory analysis shows that trained liquids have 83.67 ± 5.79 ms longer fading memory than random liquids which have shown 92.8 ± 5.03 ms fading memory for a particular type of spike train inputs. We also throw some light on the dynamics of the evolution of recurrent connections within the liquid. Moreover, compared to 'Separation Driven Synaptic Modification' - a recently proposed algorithm for iteratively refining reservoirs, our learning rule provides 9.30%, 15.21% and 12.52% more liquid separations and 2.8%, 9.1% and 7.9% better classification accuracies for four, eight and twelve class pattern recognition tasks respectively.

Title: An Online Unsupervised Structural Plasticity Algorithm for Spiking Neural Networks

Authors: S. Roy and A. Basu

Published: IEEE Transactions on Neural Networks and Learning Systems

Abstract: In this article, we propose a novel Winner-Take-All (WTA) architecture employing neurons with nonlinear dendrites and an online unsupervised structural plasticity rule for training it. Further, to aid hardware implementations, our network employs only binary synapses. The proposed learning rule is inspired by Spike-Timing-Dependent Plasticity (STDP) but differs for each dendrite based on its activation level. It trains the WTA network through formation and elimination of connections between inputs and synapses. To demonstrate the performance of the proposed network and learning rule, we employ it to solve two, four and six-class classification of random Poisson spike time inputs. The results indicate that by proper tuning of the inhibitory time constant of the WTA, a trade-off between specificity and sensitivity of the network can be achieved. We use the inhibitory time constant to set the number of subpatterns per pattern we want to detect. We show that while the percentage of successful trials are 92%, 88% and 82% for two, four and six class classification when no pattern subdivisions are made, it increases to 100% when each pattern is subdivided into 5 or 10 subpatterns. However, the former scenario of no pattern subdivision is more jitter resilient than the later ones.

Title: A Low-voltage, Low power STDP Synapse implementation using Domain-Wall Magnets for Spiking Neural Networks

Authors: G. Narasimman, S. Roy, X. Fong, K. Roy, C. H. Chang  and A. Basu

Published: IEEE International Symposium on Circuits and Systems (ISCAS), 2016

Abstract: Online, real-time learning in neuromorphic circuits have been implemented through variants of Spike-Timing-Dependent Plasticity (STDP). Current implementations have used either floating-gate devices or memristors to implement such learning synapses together with non-volatile storage. However, these approaches require high voltages ( 3-12V) for weight update and entail high energy for learning ( 4-30pJ/write). We present a domain wall memory based low-voltage, low-energy STDP synapse that can operate with a power supply as low as 0.8V and update the weight at  40fJ/write. Device level simulations are performed to prove its feasibility. Its use in associative learning is also demonstrated by using neurons with dendritic branches to classify spike patterns from MNIST dataset.

 

round_concept_cropped

2015

 

Title: Learning Spike time codes through Morphological Learning with Binary Synapses

Authors: S. Roy, P. P. San, S. Hussain, L. W. Wei and A. Basu

Published: IEEE Transactions on Neural Networks and Learning Systems

Abstract: In this brief, a neuron with nonlinear dendrites (NNLD) and binary synapses that is able to learn temporal features of spike input patterns is considered. Since binary synapses are considered, learning happens through formation and elimination of connections between the inputs and the dendritic branches to modify the structure or “morphology” of the NNLD. A morphological learning algorithm inspired by the 'Tempotron', i.e., a recently proposed temporal learning algorithm–is presented in this work. Unlike 'Tempotron', the proposed learning rule uses a technique to automatically adapt the NNLD threshold during training. Experimental results indicate that our NNLD with 1-bit synapses can obtain similar accuracy as a traditional Tempotron with 4-bit synapses in classifying single spike random latency and pair-wise synchrony patterns. Hence, the proposed method is better suited for robust hardware implementation in the presence of statistical variations. We also present results of applying this rule to real life spike classification problems from the field of tactile sensing.

Title: A Current-mode Spiking Neural Classifier with Lumped Dendritic Nonlinearity

Authors: A. Banerjee, S. Kar, S. Roy, A. Bhaduri and A. Basu

Published: IEEE International Symposium on Circuits and Systems (ISCAS), 2015

Abstract: We present the current mode implementation of a spiking neural classifier with lumped square law dendritic nonlinearity. It has been shown earlier that such a system with binary synapses can be trained with structural plasticity algorithms to achieve comparable classification accuracy with less synaptic resources than conventional algorithms. Hence, in our address event based implementation, we save 2 − 12X memory resources in storing connectivity information. The chip fabricated in 0.35μm CMOS has 8 dendrites per cell and uses two opposing cells per class to cancel common mode inputs. Preliminary results show the chip is functional and dissipates 30nW of static power per neuronal cell and 422pJ/spike.

Title: On-chip Machine Learner for Spike Sorting in Implantable Brain Machine Interfaces (BMI)

Authors: S. Korde, S. Roy, E. Yao and A. Basu

Published: IRC Conference on Science, Engineering and Technology, 2015

Abstract: Advances in neuroscience have enabled the rapid development of electronics for prostheses. The neural signals can be detected and amplified with Multi-Electrode Arrays (MEAs) of the order of 1000; and neural amplifiers, respectively. Modern day recordings use single probes for multiple neuron activity, as studying isolated cells does not present the real-life scenario. The issue of accurately identifying neural or 'spike' signals with their corresponding characteristic neurons is known as 'Spike Sorting', and it consists of a two-step process: Feature Extraction and Clustering. The motivation behind this research is to propose novel bio inspired feature extraction for the purpose of spike sorting. First, results were matched to papers that proved that derivative-based features performed better in terms of noise and error as compared to the established Principal Component Analysis (PCA). Next, a formal spiky neuron model, the Integrate-and-Fire neuron was functionally modeled on software (MATLAB) and implemented at the transistor level on Spice (CADENCE). Output firing rates, or 'Gains' of both were matched and new features were proposed from the outputs of the spiky neuron model. These features were optimized for error and showed promising results for future research in the area.

 

round_concept_cropped

2014

 

Title: Architectural exploration for on-chip, online learning in spiking neural networks

Authors: S. Roy, S. Kar and A. Basu

Published: Proceedings of the 14th International Symposium on Integrated Circuits (ISIC), 2014

Abstract: In the recent past there has been an increasing demand for area and energy efficient on-chip implementation of Machine Learning techniques. In this context we have proposed area and power optimized architectures for hardware implementation of a recently proposed supervised learning technique named Network Rewiring (NRW) for Dendritically Enhanced Readout (DER). We show that for the most optimized architecture there is a 8.5 times reduction in critical resources while the MAE has increased only by 1.76% compared to the non-optimized architecture. Moreover, for accommodating real-time training, we have also proposed an online version of the NRW rule. We also show that, though this online algorithm uses an averaging circuit having 4200 times lesser time constant compared to batch learning, yet it provides comparable performance due to the introduction of a voting mechanism.

Title: Liquid State Machine with Dendritically Enhanced Readout for Low-Power, Neuromorphic VLSI Implementations

Authors: S. Roy, A. Banerjee and A. Basu

Published: IEEE Transactions on Biomedical Circuits and Systems

Abstract: In this paper, we describe a new neuro-inspired, hardware-friendly readout stage for the liquid state machine (LSM), a popular model for reservoir computing. Compared to the parallel perceptron architecture trained by the p-delta algorithm, which is the state of the art in terms of performance of readout stages, our readout architecture and learning algorithm can attain better performance with significantly less synaptic resources making it attractive for VLSI implementation. Inspired by the nonlinear properties of dendrites in biological neurons, our readout stage incorporates neurons having multiple dendrites with a lumped nonlinearity (two compartment model). The number of synaptic connections on each branch is significantly lower than the total number of connections from the liquid neurons and the learning algorithm tries to find the best 'combination' of input connections on each branch to reduce the error. Hence, the learning involves network rewiring (NRW) of the readout network similar to structural plasticity observed in its biological counterparts. We show that compared to a single perceptron using analog weights, this architecture for the readout can attain, even by using the same number of binary valued synapses, up to 3.3 times less error for a two-class spike train classification problem and 2.4 times less error for an input rate approximation task. Even with 60 times larger synapses, a group of 60 parallel perceptrons cannot attain the performance of the proposed dendritically enhanced readout. An additional advantage of this method for hardware implementations is that the 'choice' of connectivity can be easily implemented exploiting address event representation (AER) protocols commonly used in current neuromorphic systems where the connection matrix is stored in memory. Also, due to the use of binary synapses, our proposed method is more robust against statistical variations.

 

round_concept_cropped

2013

 

Title: Hardware efficient, neuromorphic dendritically enhanced readout for liquid state machines

Authors: S. Roy, A. Basu and S. Hussain

Published: IEEE Biomedical Circuits and Systems Conference (BioCAS), 2013

Abstract: In this article, we describe a new neuro-inspired, hardware-friendly readout stage for the liquid state machine (LSM) that is suitable for on-sensor computing in resource constrained applications. Compared to the state of the art parallel perceptron readout (PPR), our readout architecture and learning algorithm can attain better performance with significantly less synaptic resources making it attractive for VLSI implementation. Inspired by the nonlinear properties of dendrites in biological neurons, our readout stage incorporates neurons having multiple dendrites with a lumped nonlinearity (two compartment model). The number of synaptic connections on each branch is significantly lower than the total number of connections from the liquid neurons and the learning algorithm tries to find the best 'combination' of input connections on each branch to reduce the error. Hence, the learning involves network rewiring (NRW) of the readout network similar to structural plasticity observed in its biological counterparts. We show that even while using binary synapses, our method can achieve 2.4 − 3.3 times less error compared to PPR using same number of high resolution synapses. Conversely, PPR requires 40−60 times more synapses to attain error levels comparable to our method.

Title: A simulated weed colony system with subregional differential evolution for multimodal optimization

Authors: S. Roy, S. M. Islam, S. Das, S. Ghosh and A. V. Vasilakos

Published: Engineering Optimization, Taylor and Francis

Abstract: This article proposes a two-stage hybrid multimodal optimizer based on invasive weed optimization (IWO) and differential evolution (DE) algorithms for locating and preserving multiple optima of a real-parameter functional landscape in a single run. Both IWO and DE have been modified from their original forms to meet the demands of the multimodal problems used in this work. A p-best crossover operation is introduced in the subregional DEs to improve their exploitative behavior. The performance of the proposed algorithm is compared with a number of state-of-the-art multimodal optimization algorithms over a benchmark suite comprising 21 basic multimodal problems and seven composite multimodal problems. Experimental results suggest that the proposed technique is able to provide better and more consistent performance over the existing well-known multimodal algorithms for the majority of test problems without incurring any serious computational burden.

Title: Multimodal optimization by artificial weed colonies enhanced with localized group search optimizers

Authors: S. Roy, S. M. Islam, S. Das and S. Ghosh

Published: Applied Soft Computing

Abstract: Multimodal optimization aims at finding multiple global and local optima (as opposed to a single solution) of a function, so that the user can have a better knowledge about different optimal solutions in the search space and as and when needed, the current solution may be switched to another suitable one while still maintaining the optimal system performance. Evolutionary Algorithms (EAs) due to their population-based approach are able to detect multiple solutions within a population in a single simulation run and have a clear advantage over the classical optimization techniques, which need multiple restarts and multiple runs in the hope that a different solution may be discovered every run, with no guarantee, however. This article proposes a hybrid two-stage optimization technique that firstly employs Invasive Weed Optimization (IWO), an ecologically inspired algorithm to find the promising Euclidean sub-regions surrounding multiple global and local optima. IWO is run for 80% of the total budget of function evaluations (FEs), and consecutively the search is intensified by using a modified Group Search Optimizer (GSO), in each detected sub-region. GSO, invoked in each sub-region discovered with IWO, is continued for 20% of the total budget of FEs. Both IWO and GSO have been modified from their original forms to meet the demands of the multimodal problems used in this work. The performance of the proposed algorithm is compared with a number of state-of-the-art multimodal optimization algorithms over a benchmark suite comprising of 21 basic multimodal problems and 7 composite multimodal problems. A practical multimodal optimization problem concerning the design of dielectric composites has also been used to test the performance of the algorithm. Experimental results suggest that the proposed technique is able to provide better and more consistent performance over the existing well-known multimodal algorithms for the majority of the test problems without incurring any serious computational burden.

Title: A Modi fied Di fferential Evolution for Symbol Detection in MIMO-OFDM System

Authors: A. Sen, S. Roy and S. Das

Published: International Conference on Swarm, Evolutionary, and Memetic Computing (SEMCCO), 2013

Abstract: It is essential to estimate the Channel and detect symbol in multiple-input and multiple-output (MIMO)-orthogonal frequency division multiplexing (OFDM) systems. Symbol detection by applying the maximum likelihood (ML) detector gives an excellent performance but in systems with a higher number of antennas and greater constellation size, the computational complexity of this algorithm becomes quite high. In this paper, we apply a recently developed modified Differential Evolution (DE) algorithm with a novel mutation, crossover as well as parameter adaptation strategies (MDE_pBX) for reducing the search space of the ML detector and the computational complexity of symbol detection in MIMO-OFDM systems. The performance of MDE_pBX has been compared with two classical symbol detectors namely ML and ZF and two famous evolutionary algorithms namely SaDE and CLPSO.

 

round_concept_cropped

2012

 

Title: An adaptive differential evolution algorithm with novel mutation and crossover strategies for global numerical optimization

Authors: S. M. Islam, S. Das, S. Ghosh, S. Roy and P. N. Suganthan

Published: IEEE Transactions on Systems, Man, and Cybernetics – Part B

Abstract: Differential evolution (DE) is one of the most powerful stochastic real parameter optimizers of current interest. In this paper, we propose a new mutation strategy, a fitness induced parent selection scheme for the binomial crossover of DE, and a simple but effective scheme of adapting two of its most important control parameters with an objective of achieving improved performance. The new mutation operator, which we call DE/current-to-gr_best/1, is a variant of the classical DE/current-to-best/1 scheme. It uses the best of a group (whose size is q% of the population size) of randomly selected solutions from current generation to perturb the parent (target) vector, unlike DE/current-to-best/1 that always picks the best vector of the entire population to perturb the target vector. In our modified framework of recombination, a biased parent selection scheme has been incorporated by letting each mutant undergo the usual binomial crossover with one of the p top-ranked individuals from the current population and not with the target vector with the same index as used in all variants of DE. A DE variant obtained by integrating the proposed mutation, crossover, and parameter adaptation strategies with the classical DE framework (developed in 1995) is compared with two classical and four state-of-the-art adaptive DE variants over 25 standard numerical benchmarks taken from the IEEE Congress on Evolutionary Computation 2005 competition and special session on real parameter optimization. Our comparative study indicates that the proposed schemes improve the performance of DE by a large magnitude such that it becomes capable of enjoying statistical superiority over the state-of-the-art DE variants for a wide variety of test problems. Finally, we experimentally demonstrate that, if one or more of our proposed strategies are integrated with existing powerful DE variants such as jDE and JADE, their performances can also be enhanced.

Title: A Differential Covariance Matrix Adaptation Evolutionary Algorithm for real parameter optimization

Authors: S. Ghosh, S. Das, S. Roy, S. M. Islam and P. N. Suganthan

Published: Information Sciences

Abstract: Hybridization in context to Evolutionary Computation (EC) aims at combining the operators and methodologies from different EC paradigms to form a single algorithm that may enjoy a statistically superior performance on a wide variety of optimization problems. In this article, we propose an efficient hybrid evolutionary algorithm that embeds the difference vector-based mutation scheme, the crossover and the selection strategy of Differential Evolution (DE) into another recently developed global optimization algorithm known as Covariance Matrix Adaptation Evolutionary Strategy (CMA-ES). CMA-ES is a stochastic method for real parameter (continuous domain) optimization of non-linear, non-convex functions. The algorithm includes an adaptation of covariance matrix which is basically an alternative method of traditional Quasi-Newton method for optimization based on gradient method. The hybrid algorithm, referred by us as Differential Covariance Matrix Adaptation Evolutionary Algorithm (DCMA-EA), turns out to possess a better blending of the explorative and exploitative behaviors as compared to the original DE and original CMA-ES, through empirical simulations. Though CMA-ES has emerged itself as a very efficient global optimizer, its performance deteriorates when it comes to dealing with complicated fitness landscapes, especially landscapes associated with noisy, hybrid composition functions and many real-world optimization problems. In order to improve the overall performance of CMA-ES, the mutation, crossover and selection operators of DE have been incorporated into CMA-ES to synthesize the hybrid algorithm DCMA-EA. We compare DCMA-EA with original DE and CMA-EA, two best known DE-variants: SaDE and JADE, and two state-of-the-art real optimizers: IPOP-CMA-ES (Restart Covariance Matrix Adaptation Evolution Strategy with increasing population size) and DMS-PSO (Dynamic Multi Swarm Particle Swarm Optimization) over a test-suite of 20 shifted, rotated, and compositional benchmark functions and two engineering optimization problems. Our comparative study indicates that although the hybridization scheme does not impose any serious burden on DCMA-EA in terms of number of Function Evaluations (FEs), DCMA-EA still enjoys a statistically superior performance over most of the tested benchmarks and especially over the multi-modal, rotated, and compositional ones in comparison to the other algorithms considered here.

Title: Adaptive IIR system identi fication using JADE

Authors: S. Roy, S. Z. Martinez and C. A. Coello Coello

Published: World Automation Congress (WAC), 2012

Abstract: Conventional mathematical programming techniques have several drawbacks related to lack of stability and premature convergence when applied to the identification of adaptive IIR systems. Additionally, such mathematical methods normally fail when reduced order adaptive models are used for the identification of higher order systems. In this paper, the IIR system identification task is formulated as an optimization problem and a variant of differential evolution (DE) called JADE is adopted to solve the problem. The explorative mutation scheme and innovative parameter adaptation schemes of JADE avoid premature convergence and increase the robustness of the DE algorithm. Both actual and reduced order identification of some benchmark IIR plants is carried out through a simulation study. The results indicate that the method adopted here is able to obtain better performance than those found by several state-of-the-art metaheuristic algorithms.

Title: A Multi-Objective Evolutionary approach for linear antenna array design and synthesis

Authors: S. Roy, S. Z. Martinez and C. A. Coello Coello

Published: IEEE Congress on Evolutionary Computation (CEC), 2012

Abstract: The linear antenna array design problem is one of the most important in electromagnetism. While designing a linear antenna array, the goal of the designer is to achieve the "minimum average side lobe level" and a "null control" in specific directions. In contrast, to the existing methods that attempt to minimize a weighted sum of these two objectives considered here, in this paper our contribution is two-fold. First, we have considered these as two distinct objectives which are optimized simultaneously in a multi-objective framework. Second, for directivity purposes, we have introduced another objective called the "maximum side lobe level" in the design formulation. The resulting multi-objective optimization problem is solved by using the recently-proposed decomposition-based Multi-Objective Particle Swarm Optimizer (dMOPSO). Our experimental results indicate that the proposed approach is able to obtain results which are better than those obtained by two other state-of-the-art Multi-Objective Evolutionary Algorithms (MOEAs). Additionally, the individual minima reached by dMOPSO outperforms those achieved by two single-objective evolutionary algorithms.

 

round_concept_cropped

2011

 

Title: A modi fied diff erential evolution for autonomous deployment and localization of sensor nodes

Authors: S. Roy, S. M. Islam, S. Ghosh and S. Das

Published: Proceedings of the 13th annual conference companion on Genetic and evolutionary computation (GECCO), 2011

Abstract: The performance of a wireless sensor network (WSN) is largely influenced by the optimal deployment and accurate localization of sensor nodes. This article considers real-time autonomous deployment of sensor nodes from an unmanned aerial vehicle (UAV). This kind of deployment has importance, particularly in ad hoc WSNs, for emergency applications, such as disaster monitoring and battlefield surveillance. The objective is to deploy the nodes only in the terrains of interest, which are identified by segmentation of the images captured by a camera onboard the UAV. In this article, we propose an improved variant of an important evolutionary algorithm Differential Evolution for image segmentation and for distributed localization of the deployed nodes. Simulation results show that the proposed algorithm ADE_pBX performs image segmentation faster than both types of algorithm for optimal thresholds. Moreover, in the case of localization, it gives more accurate results than the compared algorithms.

Title: An adaptive differential evolution algorithm for autonomous deployment and localization of sensor nodes

Authors: S. Roy, S. M. Islam, S. Ghosh, S. Das and A. V. Vasilakos

Published: Progress In Electromagnetics Research B

Abstract: In recent years, Wireless Sensor Networks (WSNs) have transitioned from being objects of academic research interest to a technology that is frequently being employed in real-life applications and rapidly being commercialized. The performance of a WSN is largely a®ected by high-quality deployment and precise localization of sensor nodes. This article deliberates autonomous deployment of sensor nodes from an Unmanned Aerial Vehicle (UAV). This kind of deployment has importance in emergency applications, such as disaster monitoring and battlefield surveillance. The goal is to deploy the nodes only in the terrains of interest, which are distinguished by segmentation of the images captured by a camera onboard the UAV. In this article, we propose an improved variant of a very powerful real parameter optimizer, called Differential Evolution (DE) for image segmentation and for distributed localization of the deployed nodes. Image segmentation for autonomous deployment and distributed localization are designed as multidimensional optimization problems and are solved by the proposed algorithm. The performance of the proposed algorithm is compared with other prominent adaptive DE-variants like SaDE and JADE as well as a powerful variant of the Particle Swarm optimization (PSO) algorithm, called CLPSO. Simulation results indicate that the proposed algorithm performs image segmentation faster than both types of algorithm for optimal thresholds. Moreover, in the case of localization, it gives more accurate results than the compared algorithms. So by using the proposed variant of Differential Evolution improvement has been achieved both in the case of speed and accuracy.

Title: Peak-to-average power ratio reduction in OFDM systems using an adaptive diff erential evolution algorithm

Authors: S. Ghosh, S. Roy, S. Das, A. Abraham and S. M. Islam

Published: IEEE Congress of Evolutionary Computation (CEC), 2011

Abstract: Orthogonal Frequency Division Multiplexing (OFDM) has emerged as very popular wireless transmission technique in which digital data bits are transmitted at a high speed in a radio environment. But the high peak-to-average power ratio (PAPR) is the major setback for OFDM systems demanding expensive linear amplifiers with wide dynamic range. In this article, we introduce a low-complexity partial transmit sequence (PTS) technique for diminishing the PAPR of OFDM systems. The computational complexity of the exhaustive search technique for PTS increases exponentially with the number of sub-blocks present in an OFDM system. So we propose a modified Differential Evolution (DE) algorithm with a novel mutation, crossover as well as parameter adaptation strategies (MDE_pBX) for a sub-optimal PTS for PAPR reduction of OFDM systems. MDE_pBX is utilized to search for the optimum phase weighting factors and extensive simulation studies have been conducted to show that MDE_pBX can achieve lower PAPR as compared to other significant DE and PSO variants like JADE, SaDE and CLPSO.

Title: A Modifi ed Discrete Di fferential Evolution based TDMA scheduling scheme for many to one communication in wireless sensor networks

Authors: S. M. Islam, S. Ghosh, S. Das, A. Abraham and S. Roy

Published: IEEE Congress of Evolutionary Computation (CEC), 2011

Abstract: Time Division Multiple Access (TDMA) plays an important role in MAC (Medium Access Control) for wireless sensor networks providing real-time guarantees and potentially reducing the delay and also it saves power by eliminating collisions. In TDMA based MAC, the sensors are not allowed to radiate signals when they are not engaged. On the other hand, if there are too many switching between active and sleep modes it will also unnecessary waste energy. In this paper, we have presented a multi-objective TDMA scheduling problem that has been demonstrated to prevent the wasting of energy discussed above and also further improve the time performance. A Modified Discrete Differential Evolution (MDDE) algorithm has been proposed to enhance the converging process in the proposed effective optimization framework. Simulation results are given with different network sizes. The results are compared with the Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) and the original Discrete DE algorithm (DDE). The proposed MDDE algorithm has successfully outperformed these three algorithms on the objective specified, which is the total time or energy for data collection.

Title: A di fferential covariance matrix adaptation evolutionary algorithm for global optimization

Authors: S. Ghosh, S. Roy, S. M. Islam, S. Das and P. N. Suganthan

Published: IEEE Symposium on Di fferential Evolution (SDE), 2011

Abstract: Covariance Matrix Adaptation Evolutionary Strategy (CMA-ES) is arguably one of the most powerful stochastic real parameter optimization algorithms in current use for non-linear non-convex functions with parameter linkages. Differential Evolution (DE) is again a very powerful but simple evolutionary algorithm for real parameter optimization. In this article, we propose a simple but very efficient hybrid evolutionary algorithm named Differential Covariance Matrix Adaptation Evolutionary Algorithm (DCMA-EA), where it creates new population members by using controlled share of its target and the population mean, the scaled difference from the current population and the step-size generated through the Covariance Matrix Adaptation. It also incorporates the selection and crossover strategies of DE. The proposed hybrid algorithm has more pronounced explorative and exploitative behaviors than its two ancestors (CMA-ES and DE). We compare DCMA-EA with original CMA-ES, some of the most known DE-variants: SaDE and JADE, and a PSO-based state-of-the-art real optimizer: DMS-PSO (Dynamic Multi-Swarm Particle Swarm optimization) and DE/Rand/1/Bin over a test-suite of 20 shifted, rotated, and compositional numerical benchmarks.

Title: Design of two-channel quadrature mirror fi lter bank: a multi-objective approach

Authors: S. Roy, S. M. Islam, S. Ghosh, S. Zhao, P. N. Suganthan and S. Das

Published: Proceedings of the 2nd International Conference on Swarm, Evolutionary, and Memetic Computing (SEMCCO), 2011

Abstract: In Digital Signal processing domain the Quadrature Mirror Filter (QMF) design problem is one of the most important problems of current interest. While designing a Quadrature Mirror Filter the goal of the designer is to achieve minimum values of Mean Square Error in Pass Band (MSEP), Mean Square Error in Stop Band (MSES), Square error of the overall transfer function of the QMF bank at the quadrature frequency and Measure of Ripple (mor). In contrast to the existing optimization-based methods that attempt to minimize a weighted sum of the four objectives considered here, in this article we consider these as four distinct objectives that are to be optimized simultaneously in a multi-objective framework. To the best of our knowledge, this is the first time to apply MO approaches to solve this problem. We use one of the best known Multi-Objective Evolutionary Algorithms (MOEAs) of current interest called NSGA-II as the optimizer. The multiobjective optimization (MO) approach provides greater flexibility in design by producing a set of equivalent final solutions from which the designer can choose any solution as per requirements. Extensive simulations reported shows that results of NSGA-II are superior to that obtained by two state-of-the-art single objective optimization algorithms namely DE and PSO.

Title: Synthesis and design of thinned planar concentric circular antenna array - a multi-objective approach

Authors: S. M. Islam, S. Ghosh, S. Roy, S. Zhao, P. N. Suganthan and S. Das

Published: Proceedings of the 2nd International Conference on Swarm, Evolutionary, and Memetic Computing (SEMCCO), 2011

Abstract: Thinned concentric antenna array design is one of the most important electromagnetic optimization problems of current interest. This antenna must generate a pencil beam pattern in the vertical plane along with minimized sidelobe level (SLL) and desired HPBW, FNBW and number of switched off elements. In this article, for the first time to the best of our knowledge, a multi-objective optimization framework for this design is presented. The four objectives described in this article are treated as four distinct objectives that are optimized simultaneously by the algorithm. The multi-objective approach provides greater flexibility by yielding a set of equivalent final solutions from which the user can choose one that attains a suitable trade-off margin as per requirements. In this article, we have used a multi-objective algorithm of current interest namely the NSGA-II algorithm. There are two types of design, one with uniform interelement spacing fixed at 0.5λ and the other with optimum uniform interelement spacing. Extensive simulation and results are given with respect to the obtained HPBW, SLL, FNBW and number of switched off elements and compared with two state-of-the-art single objective optimization methods namely DE and PSO.

Title: Non-uniform circular-shaped antenna array design and synthesis - a multi-objective approach

Authors: S. Ghosh, S. Roy, S. M. Islam, S. Zhao, P. N. Suganthan and S. Das

Published: Proceedings of the 2nd International Conference on Swarm, Evolutionary, and Memetic Computing (SEMCCO), 2011

Abstract: Design of non-uniform circular antenna arrays is one of the important optimization problems in the electromagnetic domain. While designing a non-uniform circular array, the goal of the designer is to achieve minimum side lobe levels with maximum directivity. In contrast to the single-objective methods that attempt to minimize a weighted sum of the four objectives considered here, in this article we consider these as four distinct objectives that are to be optimized simultaneously in a multi-objective (MO) framework using one of the best known Multi-Objective Evolutionary Algorithms (MOEAs) called NSGA-II. This MO approach provides greater flexibility in design by producing a set of final solutions with different trade-offs among the four objectives from which the designer can choose one as per requirements. To the best of our knowledge, other than the single objective approaches, no MOEA has been applied to design a non-uniform circular array before. Simulations have been conducted to show that the best compromise solution obtained by NSGA-II is far better than the best results achieved by the single objective approaches by using the differential evolution (DE) algorithm and the Particle Swarm Optimization (PSO) algorithm.

 

round_concept_cropped

2010

 

Title: Adaptive Di erential Evolution with p-Best Crossover for Continuous Global Optimization

Authors: S. M. Islam, S. Ghosh, S. Roy and S. Das

Published: Proceedings of the 1st International Conference on Swarm, Evolutionary, and Memetic Computing (SEMCCO), 2010

Abstract: Differential Evolution (DE) is arguably one of the most powerful stochastic real parameter optimization algorithms in current use. DE operates through the similar computational steps as employed by a standard Evolutionary Algorithm (EA). However, unlike the traditional EAs, the DE variants perturb the current-generation population members with the scaled differences of randomly selected and distinct population members. Therefore, no separate probability distribution has to be used, which makes the scheme self-organizing in this respect. Its performance, however, is still quite dependent on the setting of control parameters such as the mutation factor and the crossover probability according to both experimental studies and theoretical analyses. Our aim is to design a DE algorithm with control parameters such as the scale factor and the crossover constants adapting themselves to different problem landscapes avoiding any user intervention. Further to improve the convergence performance an innovative crossover mechanism is proposed here.

Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each authors copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.