International Journal of Bio-Inspired Computation (14 papers in press)

Regular Issues

• Recognition of crop leaf diseases based on multi-feature fusion and evolutionary algorithm optimization  
  by Lixia Zhang, Kangshun Li, Yu Qi 
  Abstract: Crop leaf disease identification refers to automatically recognize crop leaf pictures suffering from disease, to determine the type of diseases, which is important for agricultural production. Much progress has been made in this field, but there are still many challenges. For example, there are not enough ideal schemes for either disease spot area segmentation or feature representation and matching. In order to meet these challenges, a new crop leaf disease recognition method was proposed in this paper. First, disease spot segmentation method combined ultra-green feature and threshold segmentation was presented. Then, feature representation scheme with multiple features was proposed, which combined color, texture, and shape features. Finally, evolutionary algorithm was used to optimize similarity function for feature matching. Experimental results show that the scheme proposed in this paper can effectively improve recognition accuracy and has a certain practical value.
  Keywords: Crop leaf disease recognition; Evolutionary algorithm; Disease spot area segmentation; Feature representation; Feature matching.
  
  
• Bio-inspired Algorithms for Cybersecurity  
  by Kwok Tai Chui, Ryan Wen Liu, Mingbo Zhao, Xinyu Zhang 
  Abstract: It is witnessed that the popularity of the research in cybersecurity using bio-inspired algorithms (a key subset of natural algorithms) is ever-growing. As emergent research area, researchers have devoted efforts in applying and comparing various bio-inspired algorithms to the cybersecurity applications. It is in need to have a systematic review on the bio-inspired algorithms for cybersecurity to fill the gap of missing research study in this topic. The research contributions of this review article are four-fold. It first highlights the foundation of the baseline and latest development of 12 popular bio-inspired algorithms in three categories namely ecology-based, evolutionary-based and swarm intelligence-based algorithms. A systematic review is conducted to synthesise and compare the research methodologies, results and limitations. In-depth discussion will be made on the shortlisted and highly cited articles. The tips to select appropriate algorithm or the combination of multiple algorithms have been reported, along with the pros and cons on the design and formulations. Future research directions will be presented to meet the trends and unexplored research.
  Keywords: bio-inspired algorithms; cybersecurity; ecology-based algorithm; machine learning; evolutionary algorithms; multi-objective optimisation; swarm intelligence; trade-off solution.
  DOI: 10.1504/IJBIC.2023.10048934
   
  
• An algorithm of finding rules for a class of cellular automata  
  by Lei Kou, Fangfang Zhang, Luobing Chen, Wende Ke, Quande Yuan, Junhe Wan, Zhen Wang 
  Abstract: A cellular automata (CA) is an important modelling paradigm for complex systems. In the design of CA, the most difficult task is to find the transformation rules that describe the temporal evolution or pattern of a modelled system. A CA with weights (CAW) yields transition rules algorithm is proposed in this paper, which has ample physical meanings and extend the category of CA. Firstly, the weights are increased to connect the updated cell and its neighbours, and the output of each cell depends on the states of cells in the neighbourhood and their respective weights. Secondly, the error correction algorithm is adopted to find correct transition rules by adjusting weights. When the error is zero, the required transition rules with correct weights will be found to describe the fixed configuration. The CAW with the correct rules will relax to the fixed configuration regardless of the initial states. Finally, the mathematical analysis and simulation are carried out with one-dimensional CAW, and the results show that the proposed algorithm has the ability to find correct transition rules as the error converges exponentially.
  Keywords: cellular automaton with weights; CAW; transition rules; updated cells; fixed configuration.
  DOI: 10.1504/IJBIC.2022.10050857
   
  
• Neighbourhood-Based Small-World Networks Differential Evolution with Novelty Search Strategy  
  by Ying Huang, Lai Ling, Wu Huafang 
  Abstract: Differential evolution (DE) algorithms that focus too much on local search ability will lead to premature convergence, but excessive reliance on global search ability will cause slow convergence and poor search ability. Therefore, balancing global search ability and local search ability is a controversial research topic in the evolutionary process of DE. In this paper, we propose a DE algorithm based on small-world network topology and novelty search (NSWDE), which uses small-world network to construct dynamic neighborhoods and form structured populations, changing the dynamics of information exchange within populations. The use of neighborhoods enables the evolutionary algorithm to better explore the search space. A novelty search strategy that will continuously search for new behaviors in the search space is introduced. Novelty search tends to identify more evolutionary solutions than traditional target search. Additionally, we set different scaling factors for the two different drivers based on fitness and novelty, and modulated the weights of the two scaling factors by using the different characteristics, which can effectively regulate the population diversity.
  Keywords: differential evolution algorithm; small world network; novelty search; topology structure; parameter adaptation.
  DOI: 10.1504/IJBIC.2022.10054188
   
  
• Chinese machine reading comprehension based on deep learning neural network  
  by Chao Ma, Jing An, Jing Xu, BinChen Xu, Luyuan Xu, Xiang-En Bai 
  Abstract: Most of the existing MRC (Machine Reading Comprehension) models mainly learn the attention mechanism of interactive alignment between questions and texts, and there are still great difficulties in the characterization of ultra-long texts. The complex grammar, diverse expressions, and flexible structure of the Chinese corpus make the MRC task even more difficult. In this paper, a Transformer-based neural network model XBCNet (eXtremely Brisk and Clever Net) was proposed for the MRC problem of Chinese text. XBCNet expands the receptive field by introducing a stack of dilated convolution neural networks, and it is also closely coupled with the word vector algorithm. This process aims to reduce the training and inference time of the model, while alleviating the drawback of long questions that are difficult to answer accurately. The experimental results demonstrate that XBCNet improves the performance of MRC under general texts and obtains the best inference effect in limited computing resources.
  Keywords: natural language processing; Chinese machine reading comprehension; dilated convolution; attention mechanism; deep learning; neural networks.
  DOI: 10.1504/IJBIC.2022.10054196
   
  
• Chaotic resonance in discrete fractional-order LIF neural network motifs  
  by Hao Yin, Jiacheng Tang, Bo Liu, Shuaiyu Yao, Qi Kang 
  Abstract: Chaotic resonance (CR) is a phenomenon that nonlinear systems enhanced respond to weak signals under the influence of chaotic signals. It exists robustly in nature, including the human nervous system. Can we build a neural network model that can detect weak signals with multiple frequencies under chaotic signals? Note that fractional calculus can naturally capture intrinsic phenomena in complex dynamical. We first introduced fractional calculus and proposed the discrete fractional-order LIF model. The Triple- neuron Feed-forward Loop network motifs are also established. The proposed model has rich response characteristics and can better detect weak signals of various frequencies in the environment. The experimental results show that neuron and neural network motifs can independently respond to a weak signal with a certain frequency by adjusting the fractional order, and network motifs can achieve orderly cluster discharge, which provides a new idea for us to build deeper spiking neural networks and explore the mechanism of weak signal detection and transmission in biological nervous systems.
  Keywords: Fractional-order systems; neural network motifs; discrete LIF model; Chaotic resonance; dynamic behavior.
  DOI: 10.1504/IJBIC.2023.10054455
   
  
• Research on Emergence Mechanism of Collective Intelligence from the Complexity Perspective  
  by Renbin Xiao, Zhenhui Feng, Bowen Wu 
  Abstract: This paper explores the emergence mechanism of collective intelligence (CI) from the complexity perspective. It begins with a comparison of the main features based on the two basic stages of CI, i.e., CI 1.0 (swarm intelligence) and CI 2.0 (crowd intelligence). Considering the connection mechanism between the two stages is still unclear, we would regard higher organism group behaviours as the transition between lower organism group behaviours to crowd behaviours. Accordingly, the bionic prototypes of CI can be classified into three categories: lower organisms, higher organisms and humans. This paper first refined the emergence mechanisms of CI in lower organisms represented by labour division, i.e., stimulus-response mechanism and activation-inhibition mechanism. Subsequently, the higher organism emergence mechanism was revealed, which is the attraction-repulsion mechanism based on roles division and perception driven. Furthermore, the emergence mechanism of crowd intelligence at the perceptual level and cognitive level are presented respectively, by means of process evolutionary description based on the attraction-repulsion mechanism. Finally, the research gives a holistic illustration of the emergence mechanism of CI.
  Keywords: collective intelligence; emergence; complexity; stimulus-response; activation-inhibition; attraction-repulsion; perceptual level; cognitive level.
  DOI: 10.1504/IJBIC.2023.10054716
   
  
• Intelligent Agents Path Planning in Wireless Sensor Networks based on Vor-PSO Algorithm  
  by Zining Yan, Guisheng Yin, Sizhao Li, Zechao Liu 
  Abstract: Wireless sensor networks (WSN) have wide applications in various fields, and intelligent agents have been applied to different tasks in the area of WSN because of their robust monitoring and exploration ability. This work proposes a path planning scheme in WSN based on computational geometry and particle swarm optimisation (PSO), which aims to minimise the cost of intelligent agents. First, we reconstruct the WSN topology according to the Voronoi diagram, which can convert the uncertain path cost into a deterministic expression. Then, using the discrete variational method, we construct an optimal path cost function considering exposure and length. Next, we develop a path planning algorithm based on Voronoi topology and PSO (Vor-PSO). Furthermore, we design a fitness function that considers the angle between the Voronoi edge and the optimal extreme point to update the particle position. The proposed heuristic algorithm effectively solves the problem of finding a feasible path in a high-coverage WSN and can be applied to different types of WSNs and multi-agent cluster task planning. Finally, simulation results are given to prove the effectiveness and computational performance of the proposed Vor-PSO algorithm.
  Keywords: Voronoi; intelligent vehicles; path planning; particle swarm optimisation; PSO.
  DOI: 10.1504/IJBIC.2023.10054855
   
  
• Two new selection methods and their effects on the performance of genetic algorithm in solving supply chain and traveling salesman problems  
  by Sadegh Eskandari, Marjan Kuchaki Rafsanjani 
  Abstract: Genetic algorithm (GA) is a well-known evolutionary optimisation method in various operational research areas. Selection is an important operator in GA that provides a trade-off between exploitation and exploration aspects of genetic algorithm. In this paper, two new combinational selection methods called generational sequential mixed selection (GSMS) and generational random mixed selection (GRMS) are presented and compared with six existing selection operators, applied to supply chain network (SCN) design and travelling salesman problems (TSP). The experiments show that the proposed operators achieve better results than existing operators, in every way. Moreover, several state-of-the-art methods are compared with genetic algorithm versions, which adopt the proposed operators. The results on 15 TSPs show that our approach is superior in ten cases. Moreover, the results on ten SCN instances show the superiority of the proposed approach in 50% of the cases.
  Keywords: genetic algorithms; GA; selection operators; supply chain network; SCN; travelling salesman problems; TSP.
  DOI: 10.1504/IJBIC.2023.10054864
   
  
• Leveraging Knowledge Graph for Domain-Specific Chinese Named Entity Recognition via Lexicon-Based Relational Graph Transformer  
  by Yunbo Gao, Guanghong Gong, Bipeng Ye, Xingyu Tian, Ni Li, Haitao Yuan 
  Abstract: Leveraging knowledge graphs (KGs) has been an emerging direction to improve the performance of deep learning-based Chinese named entity recognition (CNER). Nevertheless, most existing methods directly inject correlated words into sentences but ignore word boundaries that are crucial for CNER. Conflicts among incorrect word segmentations may misguide models to predict incorrect labels. To solve this problem, this work investigates a novel lexicon-based relational graph transformer (LRGT), which combines relational graph-structured inputs and transformer tailored for lexicon-augmented CNER. In LRGT, characters and self-matched lexicon words are fully interacted through a two-phase relational graph softmax message passing mechanism. The finally enhanced character representation in LRGT dynamically integrates both lexical and relative positional information, which is distinguishable for the identification. Results on four benchmark datasets demonstrate that LRGT significantly outperforms several state-of-the-art methods. We further demonstrate that LRGT with KG achieves higher performance on two public specific-domain CNER datasets. LRGT performs up to 3.35 times faster than several typical baselines while achieving better F1-score by up to 1.92% and 2.24%, respectively.
  Keywords: deep learning; knowledge graph; Chinese named entity recognition; CNER; lexicon augmentation; relational graph transformer; RGT; lexicon-based relational graph transformer; LRGT.
  DOI: 10.1504/IJBIC.2023.10055548
   
  
• An Empirical Study of Improved Ant Colony Clustering Algorithm in English Composition Review  
  by Xiao Chang, Jianguang Sun 
  Abstract: The scoring analysis method of English composition review lacks flexibility. To solve this problem, this paper proposes an analysis method based on the improved ant colony clustering algorithm. Where, cosine distance and Euclidean distance were combined to determine the conversion function. The empirical results show that compared with the previous standard ant colony clustering algorithm, the traditional k-means algorithm and IGKA algorithm, the improved ant colony clustering algorithm can realise the comprehensive evaluation of English composition review. It can be seen that the proposed method is reasonable and feasible, which can effectively conduct cluster analysis on English composition review, and has a higher accuracy rate of 89.33%. Therefore, in order to achieve the clustering analysis of English composition rating more precisely, the next step is to improve the ant colony clustering algorithm by repeated experiments on experimental data.
  Keywords: ant colony clustering algorithm; cluster analysis; English composition review; score analysis.
  DOI: 10.1504/IJBIC.2023.10055786
   
  
• On the Effect of Particle Update Modes in Particle Swarm Optimization  
  by Dong Nanjiang, Rui Wang, Tao Zhang, Junwei Ou 
  Abstract: Particle swarm optimisation has been successfully applied in various single- and multi-objective optimisation problems. Through the literature review, it is shown that in PSO-based algorithms particles are updated mainly in two different modes. Specifically, the first mode denoted as PSO-a uses random vectors in [0, 1]n in the particle update process. The second mode denoted as PSO-b uses random variables in [0, 1]. This study systematically analyzed the effect of different modes on a varied set of benchmarks. Experimental results show that the PSO-a mode is more suitable for single-objective optimisation while the PSO-b has certain advantages for multi-objective optimisation due to the regularity of multi-objective problems. Also, the introduction of a mutation operator into PSO-b can overcome the limit of dimension. Moreover, to guarantee finding the optimal solution, the swarm size must be larger than the problem dimensionality when PSO-b is purely adopted.
  Keywords: evolutionary computation; particle swarm optimisation; PSO; population size; multi-objective optimisation.
  DOI: 10.1504/IJBIC.2023.10056269
   
  
• Research on a bionic swarm intelligence algorithm and model construction of the integrated dispatching system for the rescue of disaster victims  
  by Fyu Wang, Jiawen Fan, Mengkai Chen, Haoxuan Xie, Juma Nzige, Weining Li 
  Abstract: Most of the existing emergency rescue studies focus on the shortest rescue path rather than the shortest time within a reasonable rescue path, and it is only a simple resource sequencing in surgical scheduling. In this paper, casualty emergency rescue vehicle scheduling and operation scheduling are considered to be complex integrated systems. In the early stage of integration, the single side fuzzy time window model of the casualty emergency rescue vehicle is constructed according to the feature. At later stage, considering the learning-forgetting effect of physician, the mathematical model of flexible flow scheduling in the operation process of the wounded was established. This paper applies the multi-population thought to improve the basic firefly algorithm. The simulation results show that the proposed model and algorithm effectively solve the scheduling problem of the complex system of casualty rescue after disaster, which is helpful for improving the operation management system of disaster relief.
  Keywords: emergency rescue; vehicle route optimisation; surgical scheduling; bionics optimisation algorithm.
  DOI: 10.1504/IJBIC.2023.10056270
   
  
• Improved slime mould algorithm by perfecting bionic-based mechanisms  
  by Tianyu Yu, Jiawen Pan, Qian Qian, Miao Song, Jibin Yin, Yong Feng, Yunfa Fu, Yingna Li 
  Abstract: Slime mould algorithm (SMA) is a new meta-heuristic algorithm which imitates the biological mechanism of natural creatures. It has good initial performance, but it also has some disadvantages. More importantly, the bionic modelling of SMA is not complete, and many biological mechanisms of slime moulds are ignored. This paper proposes an improved slime mould algorithm by perfecting bionic mechanism (IBSMA). Specifically, three mechanisms are added. Among them, the polar growth mechanism is used to improve the global optimisation ability, the memory mechanism is used to enhance the ability of the algorithm to jump out of the local optimum, and the amoeba mechanism is used to expand the search space and improve the exploration capability of the algorithm. Qualitative and effectiveness analyses are conducted, and the proposed algorithm is compared with nine excellent algorithms. The results show that IBSMA has the best performance, which is also verified by non-parametric statistical methods.
  Keywords: slime mould algorithm; meta-heuristic algorithm; bionic modelling.
  DOI: 10.1504/IJBIC.2023.10056520