Chaotic maps play an important role in improving evolutionary algorithms (EAs) for avoiding the local optima and speeding up the convergence. However, different chaotic maps in different phases have different effects on EAs. This paper focuses on exploring the effects of chaotic maps and giving comprehensive guidance for improving multiobjective evolutionary algorithms (MOEAs) by series of experiments. NSGA‐II algorithm, a representative of MOEAs using the nondominated sorting and elitist strategy, is taken as the framework to study the effect of chaotic maps. Ten chaotic maps are applied in MOEAs in three phases, that is, initial population, crossover, and mutation operator. Multiobjective problems (MOPs) adopted are ZDT series problems to show the generality. Since the scale of some sequences generated by chaotic maps is changed to fit for MOPs, the correctness of scaling transformation of chaotic sequences is proved by measuring the largest Lyapunov exponent. The convergence metric γ and diversity metric Δ are chosen to evaluate the performance of new algorithms with chaos. The results of experiments demonstrate that chaotic maps can improve the performance of MOEAs, especially in solving problems with convex and piecewise Pareto front. In addition, cat map has the best performance in solving problems with local optima.

Particle swarm optimization (PSO) is a population-based stochastic algorithm modeled on the social behaviors observed in flocking birds. Over the past quarter century, the particle swarm optimization algorithm has attracted many researchers’ attention. Through the convergent operation and divergent operation, individuals in PSO group and diverge in the search space/objective space. In this paper, the historical development, the state-of-the-art, and the applications of the PSO algorithms are reviewed. In addition, the characteristics and issues of the PSO algorithm are also discussed from the evolution and learning perspectives. Every individual in the PSO algorithm learns from itself and another particle with good fitness value. The search performance and convergence speed were affected by different learning strategies. The scheduling and data-mining problems are illustrated as two typical cases of PSO algorithm solving real-world application problems. With the analysis of different evolution and learning strategies, particle swarm optimization algorithm could be utilized on solving more real-world application problems effectively, and the strength and limitation of various PSO algorithms could be revealed.

Outlier detection is a very essential problem in a variety of application areas. Many detection methods are deficient for high-dimensional time series data sets containing both isolated and assembled outliers. In this paper, we propose an Outlier Detection method based on Cross-correlation Analysis (ODCA). ODCA consists of three key parts. They are data preprocessing, outlier analysis, and outlier rank. First, we investigate a linear interpolation method to convert assembled outliers into isolated ones. Second, a detection mechanism based on the cross-correlation analysis is proposed for translating the high-dimensional data sets into 1-D cross-correlation function, according to which the isolated outlier is determined. Finally, a multilevel Otsu’s method is adopted to help us select the rank thresholds adaptively and output the abnormal samples at different levels. To illustrate the effectiveness of the ODCA algorithm, four experiments are performed using several high-dimensional time series data sets, which include two small-scale sets and two large-scale sets. Furthermore, we compare the proposed algorithm with the detection methods based on wavelet analysis, bilateral filtering, particle swarm optimization, auto-regression, and extreme learning machine. In addition, we discuss the robustness of the ODCA algorithm. The statistical results show that the ODCA algorithm is much better than existing mainstream methods in both effectiveness and time complexity.

In order to solve the problem of slow convergence rate and long planned path when the robot plans a path in unknown environment by using Q-learning algorithm, we propose an Experience-Memory Q-Learning (EMQL) algorithm based on the continuous update of the shortest distance from the current state node to the start point. The autonomous learning ability of the robot is enhanced by the different role assignment of two tables in the proposed algorithm. EM table with (m * 1) dimension is designed to record the distance information, reflecting the learning process of the robot. Q table is adopted as an auxiliary guidance for the experience transfer strategy and experience reuse strategy, and these strategies enable the robot accomplish the task even if the destination is changed or the path is blocked. Further, the learning efficiency of the robot in the EMQL algorithm is improved by the dual reward mechanism consisting of static reward and dynamic reward. The static reward is designed to prevent the robot from exploring a state node excessively. The dynamic reward is responsible for helping the robot avoid searching blindly in unknown environment. We verify the effectiveness of the proposed algorithm on the grid maps and road network maps. The comparison about planning time, iteration times and path length show that the performance of the EMQL algorithm is superior to Q-learning algorithm in convergence speed and optimization ability. Additionally, the practicability of the proposed algorithm is further validated in a real-world experiment using the Turtlebot3 burger robot.

Bearing is a failure-prone component in rotating machinery equipment, and various task conditions make bearing fault diagnosis (BFD) a challenging task. Despite the great success of convolutional neural network (CNN) in feature extraction, most of them are based on manual tuning, and it is not easy to determine a unified CNN architecture with an effective balance between accuracy and computational resources [parameters and floating point operations (FLOPs)]. In this article, an automatic CNN architecture design method called AUTO-CNN is proposed to address the above issues. First, a novel representation learning method (RLM) based on spatial–temporal tensors (STTs) is proposed, in which AUTO-CNN can use 2-D multiscale convolutions to capture more feature hierarchy, i.e., temporal variabilities and spatial characteristics. Second, unlike most of the current nontask-specific studies that establish deeper and more complex networks for better performance, this article investigates lightweight CNNs with limited resources under task-specific applications. Finally, to speed up the design process, an efficient hierarchical encoding strategy and space exploration strategy are proposed to avoid inefficient architecture generation and reduce CNNs training time. Experiments on four real-world BFD tasks demonstrate that the proposed method outperforms the peer competitors, including machine learning, deep learning, and other CNN-based methods.