Gaussian Perturbation Specular Reflection Learning and Golden-Sine-Mechanism-Based Elephant Herding Optimization for Global Optimization Problems

Elephant herding optimization (EHO) has received widespread attention due to its few control parameters and simple operation but still suffers from slow convergence and low solution accuracy. In this paper, an improved algorithm to solve the above shortcomings, called Gaussian perturbation specular reflection learning and golden-sine-mechanism-based EHO (SRGS-EHO), is proposed. First, specular reflection learning is introduced into the algorithm to enhance the diversity and ergodicity of the initial population and improve the convergence speed. Meanwhile, Gaussian perturbation is used to further increase the diversity of the initial population. Second, the golden sine mechanism is introduced to improve the way of updating the position of the patriarch in each clan, which can make the best-positioned individual in each generation move toward the global optimum and enhance the global exploration and local exploitation ability of the algorithm. To evaluate the effectiveness of the proposed algorithm, tests are performed on 23 benchmark functions. In addition, Wilcoxon rank-sum tests and Friedman tests with 5% are invoked to compare it with other eight metaheuristic algorithms. In addition, sensitivity analysis to parameters and experiments of the different modifications are set up. To further validate the effectiveness of the enhanced algorithm, SRGS-EHO is also applied to solve two classic engineering problems with a constrained search space (pressure-vessel design problem and tension-/compression-string design problem). The results show that the algorithm can be applied to solve the problems encountered in real production.

Multiconstrained Ascent Trajectory Optimization Using an Improved Particle Swarm Optimization Method

This paper researches the ascent trajectory optimization problem in view of multiple constraints that effect on the launch vehicle. First, a series of common constraints that effect on the ascent trajectory are formulated for the trajectory optimization problem. Then, in order to reduce the computational burden on the optimal solution, the restrictions on the angular momentum and the eccentricity of the target orbit are converted into constraints on the terminal altitude, velocity, and flight path angle. In this way, the requirement on accurate orbit insertion can be easily realized by solving a three-parameter optimization problem. Next, an improved particle swarm optimization algorithm is developed based on the Gaussian perturbation method to generate the optimal trajectory. Finally, the algorithm is verified by numerical simulation.

Adaptive Particle Swarm Optimization with Gaussian Perturbation and Mutation

The particle swarm optimization (PSO) is a wide used optimization algorithm, which yet suffers from trapping in local optimum and the premature convergence. Many studies have proposed the improvements to address the drawbacks above. Most of them have implemented a single strategy for one problem or a fixed neighborhood structure during the whole search process. To further improve the PSO performance, we introduced a simple but effective method, named adaptive particle swarm optimization with Gaussian perturbation and mutation (AGMPSO), consisting of three strategies. Gaussian perturbation and mutation are incorporated to promote the exploration and exploitation capability, while the adaptive strategy is introduced to ensure dynamic implement of the former two strategies, which guarantee the balance of the searching ability and accuracy. Comparison experiments of proposed AGMPSO and existing PSO variants in solving 29 benchmark functions of CEC 2017 test suites suggest that, despite the simplicity in architecture, the proposed AGMPSO obtains a high convergence accuracy and significant robustness which are proven by conducted Wilcoxon’s rank sum test.

Geographically Masking Addresses to Study COVID-19 Clusters

Background: The spatio-temporal analysis of cases is a good way an epidemic, and the recent COVID-19 pandemic unfortunately generated a huge amount of data. But analysing this raw data, with for instance the address of the people who contracted COVID-19, raises some privacy issues, and geomasking is necessary topreserve both people privacy and the spatial accuracy required for analysis. This paper proposes dierent geomasking techniques adapted to this COVID-19 data.Methods: Different techniques are adapted from the literature, and tested on a synthetic dataset mimicking the COVID-19 spatio-temporal spreading in Paris and a more rural nearby region. Theses techniques are assessed in terms of k-anonymity and cluster preservation.Results: Three adapted geomasking techniques are proposed: aggregation, bimodal gaussian perturbation, and simulated crowding. All three can be useful in different use cases, but the bimodal gaussian perturbation is the overall best techniques, and the simulated crowding is the most promising one, provided some improvements are introduced to avoid points with a low k-anonymity.Conclusions: It is possible to use geomasking techniques on addresses of people who caught COVID-19, while preserving the important spatial patterns.

Regularized Training and Tight Certification for Randomized Smoothed Classifier with Provable Robustness

Recently smoothing deep neural network based classifiers via isotropic Gaussian perturbation is shown to be an effective and scalable way to provide state-of-the-art probabilistic robustness guarantee against ℓ2 norm bounded adversarial perturbations. However, how to train a good base classifier that is accurate and robust when smoothed has not been fully investigated. In this work, we derive a new regularized risk, in which the regularizer can adaptively encourage the accuracy and robustness of the smoothed counterpart when training the base classifier. It is computationally efficient and can be implemented in parallel with other empirical defense methods. We discuss how to implement it under both standard (non-adversarial) and adversarial training scheme. At the same time, we also design a new certification algorithm, which can leverage the regularization effect to provide tighter robustness lower bound that holds with high probability. Our extensive experimentation demonstrates the effectiveness of the proposed training and certification approaches on CIFAR-10 and ImageNet datasets.