Obaid Ur Rehman

A modified quantum-based particle swarm optimization for engineering inverse problem

The purpose of this paper is to explore the potential of standard quantum-based particle swarm optimization (QPSO) methods for solving electromagnetic inverse problems.,A modified QPSO algorithm is designed.,The modified QPSO algorithm is an efficient and robust global optimizer for optimizing electromagnetic inverse problems. More specially, the experimental results as reported on different case studies demonstrate that the proposed method can find better final optimal solution at an early stage of the iterating process (uses less iterations) as compared to other tested optimal algorithms.,The modifications include the design of a new position updating formula, the introduction of a new mutation strategy and a dynamic control parameter to intensify the convergence speed of the algorithm.

Optimization of CNN through Novel Training Strategy for Visual Classification Problems

The convolution neural network (CNN) has achieved state-of-the-art performance in many computer vision applications e.g., classification, recognition, detection, etc. However, the global optimization of CNN training is still a problem. Fast classification and training play a key role in the development of the CNN. We hypothesize that the smoother and optimized the training of a CNN goes, the more efficient the end result becomes. Therefore, in this paper, we implement a modified resilient backpropagation (MRPROP) algorithm to improve the convergence and efficiency of CNN training. Particularly, a tolerant band is introduced to avoid network overtraining, which is incorporated with the global best concept for weight updating criteria to allow the training algorithm of the CNN to optimize its weights more swiftly and precisely. For comparison, we present and analyze four different training algorithms for CNN along with MRPROP, i.e., resilient backpropagation (RPROP), Levenberg–Marquardt (LM), conjugate gradient (CG), and gradient descent with momentum (GDM). Experimental results showcase the merit of the proposed approach on a public face and skin dataset.

A modified PSO algorithm with dynamic parameters for solving complex engineering design problem

ABSTRACT This paper proposed a new approach of particle swarm optimization (PSO). The proposed modified PSO algorithm is equipped with some specially designed mechanisms of adaptively updating algorithm parameters to preserve the diversity of the swarm and to keep the balance between exploration and exploitation searches. All these mechanisms help the algorithm to avoid the premature convergence and to strengthen its robustness. Experiments are conducted on different complicated, unimodal and multimodal test functions, as well as a typical engineering inverse problem, the TEAM Workshop problem 22. The numerical results illustrate that the proposed PSO shows better performance as compared to other well developed evolutionary algorithms.

An improved quantum based particle swarm optimizer applied to electromagnetic optimization problems

Purpose The aim of this paper is to explore the potential of standard quantum particle swarm optimization algorithms to solve single objective electromagnetic optimization problems. Design/methodology/approach A modified quantum particle swarm optimization (MQPSO) algorithm is designed. Findings The MQPSO algorithm is an efficient and robust global optimizer for optimizing electromagnetic design problems. The numerical results as reported have demonstrated that the proposed approach can find better final optimal solution at an initial stage of the iterating process as compared to other tested stochastic methods. It also demonstrates that the proposed method can produce better outcomes by using almost the same computation cost (number of iterations). Thus, the merits or advantages of the proposed MQPSO method in terms of both solution quality (objective function values) and convergence speed (number of iterations) are validated. Originality/value The improvements include the design of a new position updating formula, the introduction of a new selection method (tournament selection strategy) and the proposal of an updating parameter rule.

A dynamic particle swarm optimization method applied to global optimizations of engineering inverse problem

Purpose The aim of this paper is to explore the potential of particle swarm optimization (PSO) algorithm to solve an electromagnetic inverse problem. Design/methodology/approach A modified PSO algorithm is designed. Findings The modified PSO algorithm is a more stable, robust and efficient global optimizer for solving the well-known benchmark optimization problems. The new mutation approach preserves the diversity of the population, whereas the proposed dynamic and adaptive parameters maintain a good balance between the exploration and exploitation searches. The numerically experimental results of two case studies demonstrate the merits of the proposed algorithm. Originality/value Some improvements, such as the design of a new global mutation mechanism and introducing a novel strategy for learning and control parameters, are proposed.

An improved quantum particle swarm optimization applied to inverse problem in electromagnetics

The development of global optimal techniques for inverse problems in electromagnetics has been successful in the last few years. However, inspired from the classical Particle Swarm Optimization (PSO) algorithm and quantum mechanics, this work presents an improved Quantum based particle swarm optimization (QPSO) by using a tournament selection strategy. Also, a new index, called Tbest (tournament best), is incorporated into the QPSO to further enhance its performance. The feasibility and merit of the proposed approach are verified by mathematic functions and an electromagnetic inverse problem.

A particle swarm optimization method applied to global optimization of inverse problem

Particle Swarm Optimization (PSO) is a global optimal algorithm based on swarm intelligence. PSO is more popular due to its easiness in implementation and fast convergence speed. However, PSO will be trapped into local optima while it is used to solve complex, multimodal inverse problems. To establish a proper balance between the exploration and exploitation searches, this paper introduces a dynamic and adaptive mechanism for the three basic parameters. To preserve the diversity of the swarm, a particular best particle, which takes part in a mutation operation, is introduced in the modified PSO. The numerical results on a case study show that the proposed PSO finds the best results as compared to other ones.