International Journal of Artificial Intelligence and Soft Computing (5 papers in press)

Regular Issues

• Integrated AWA fitness PSO-SPICE framework for automated design and optimisation of analogue and mixed-signal ICs  
  by Harsha Maddur Venkataswamy, B.P. Harish 
  Abstract: The design and optimisation of analogue and mixed-signal integrated circuits become intractable with technology scaling. It gives rise to multi-dimensional tradeoffs among its numerous performance metrics. Evolutionary algorithms are being explored to generate possible solutions having goodness of fit with the desired solution. In this direction, a novel fitness evaluation function integrated with PSO and PSO-SPICE framework is proposed to design and implement multi-objective optimisation for analogue and mixed-signal circuit design automation. The framework is demonstrated to automatically design and optimise the multi-objectives of 2-stage op-amp and 4-bit flash ADC. The proposed fitness evaluation function demonstrate large design outperformance independent of quality of initial population and requiring no adaptive weights. The novel fitness function driven PSO-SPICE framework exemplifies a robust, scalable, and precise method for multi-objective optimisation of analogue and mixed-signal circuits of varying scale and design complexity.
  Keywords: particle swarm optimisation; PSO; 2-stage op-amp; flash ADC; multi-objective optimisation; AWA fitness function.
  DOI: 10.1504/IJAISC.2024.10065061
   
  
• Genetic whale optimisation algorithm for solving travelling salesman problem  
  by Amit Kumar 
  Abstract: Travelling salesman problem (TSP) is a hard combinatorial optimisation problem that has an enormous discrete search space with an excess of potential solutions. In this condition, it is impossible to carry out an exhaustive search using merely brute force. Whale optimisation algorithm (WOA) is a recent nature-inspired metaheuristic algorithm that is widely being utilised for the modern intelligent solution approach for hard optimisation problems. It is inspired by the spiral bubble-net hunting strategy of humpback whales. In this paper, a new discrete genetic operators-based whale optimisation algorithm (GWOA) has been presented for addressing the TSP. Further, experiments-based comparison of the GWOA with some recently proposed discrete particle swarm optimisation algorithms shows that the former is able to find better quality tours for TSP.
  Keywords: travelling salesman problem; TSP; nature-inspired metaheuristic algorithms; whale optimisation algorithm; WOA; particle swarm optimisation; PSO.
  DOI: 10.1504/IJAISC.2024.10064270
   
  
• Rotor fault characterisation in induction motors under different load levels via machine learning methods  
  by Hayri Arabacı, Mücahid Barstuğan 
  Abstract: Induction motors stand out for their robustness and are widely used in the industrial sector. Literature studies have focused more on rotor faults because rotor fault signatures are hard to detect. In most experimental studies, tests were carried out using a single motor for fault classification. In general, training and fault classification was conducted on a single load type. This study focused on fault classification for induction motors with varying powers and load conditions. Motor current data for four different induction motors and randomly selected load levels were obtained, a classifier structure was formed using machine learning, and tests were carried out. Classification results for the five classifiers were obtained and compared to determine the reliability of the generalised classifier structure. Support vector machines and k-nearest neighbour methods were used in the classification and k-nearest neighbour achieved at 99.51% accuracy.
  Keywords: fault classification; induction motor; machine learning; pattern recognition; rotor faults.
  DOI: 10.1504/IJAISC.2024.10064252
   
  
• Optimisation of spatial-exploitation CNN models through hyperparameter-tuning and human-in-the-loop combination  
  by Luke Beveridge, Keshav Dahal 
  Abstract: Spatial-exploitation convolutional neural networks (CNNs) have a simplified architecture compared to other CNN models. However, devices with limited computational resources could struggle with processing spatial-exploitation CNNs. To address this, we investigate two methods to optimise spatial-exploitation CNN models for time efficiency and classification accuracy: hyperparameter-tuning, and human-in-the-loop (HITL). We apply grid-search to optimise the hyperparameter space, whilst HITL is used to identify whether the time-to-accuracy relationship of the optimised model can be improved. To show the versatility of combining the two methods, CIFAR-10, MNIST, and Imagenette are used as model input. This paper contributes to spatial-exploitation CNN optimisation by combining hyperparameter-tuning and HITL. Results show that this combination improves classification accuracy by 1.47-2.34% and reduces the time taken to conduct this task by 27-28%, depending on dataset. We conclude that combining hyperparameter-tuning and HITL are a viable approach to optimise spatial-exploitation CNNs for devices with limited computational resources.
  Keywords: deep learning; convolutional neural network; CNN; image classification; hyperparameter-tuning; human-in-the-loop; HITL.
  DOI: 10.1504/IJAISC.2024.10064253
   
  
• Improved moth flame optimisation succored FOPID controller of integrated industrial processes with delay time  
  by R. Anuja, Mary A.G. Ezhil, M. Dhiviya Nycil 
  Abstract: The control of industrial process in the presence of delay time is a demanding benchmark control issue. The conventional feedback controllers and tuning approaches do not provide the controller with enough robustness to prevent delay time. Using a conventional control scheme is ineffective, and so the fractional order proportional integral derivative controller (FOPIDC) is preferred. This paper focuses an efficient optimisation control for industrial processes with delay time. The proposed technique is the execution of the improved moth flame optimisation (IMFO) scheme. This paper concentrated on three main contributions to the optimal tuning of the parameters of the FOPID controller, such as crow search optimisation (CSO) control, whale algorithm optimisation (WAO) and the proposed IMFO control. The performances of the methods are evaluated based on transient responses, convergence rate and bode plot-based stability analysis to demonstrate the effectiveness of the proposed IMFO technique in enhanced system operation.
  Keywords: delay time; tuning of FOPID; IMFO algorithm; crow search optimisation; CSO; algorithm; whale algorithm optimisation; WAO algorithm.
  DOI: 10.1504/IJAISC.2024.10064785