基于动态权重的Adaboost算法研究 *

针对Adaboost算法只能静态分配基分类器权重,不能自适应地对每个测试样本动态调整权重的问题,提出了一种基于动态权重的Adaboost算法。算法通过对训练样本集合进行聚类,并分析每个基分类器和每个类簇的适应性,进而为每个基分类器在不同类簇上设置不同权重,最终根据测试样本与类簇之间的相似性来计算基分类器在测试样本上的权重。在UCI数据集上的实验结果表明本文提出算法有效利用了测试样本之间的差异性,得到了比Adaboost算法更好的效果。     

Chaos-based multi-objective immune algorithm with a fine-grained selection mechanism

In this paper, we propose a chaos-based multi-objective immune algorithm (CMIA) with a fine-grained selection mechanism based on the clonal selection principle. Taking advantage of the ergodic and stochastic properties of chaotic sequence, a novel mutation operator, named as chaos-based mutation (CM) operator, is proposed. Moreover, the information of diversity estimation is also adopted in the CM operator for nondominated solutions to adjust mutation steps adaptively, which encourages searching less-crowded regions with relative large step sizes. When comparing with polynomial mutation operator that is used in many state-of-the-art multi-objective optimization evolutionary algorithms, simulations show that it is effective to enhance the search performance. On the other hand, in order to increase the population diversity, a fine-grained selection mechanism is proposed in this paper, which seems to be remarkably effective in two-objective benchmark functions. When comparing with two state-of-the-art multi-objective evolutionary algorithms (NSGA-II and SPEA-2) and a new multi-objective immune algorithm (NNIA), simulation results of CMIA indicate the effectiveness of the fine-grained selection mechanism and the remarkable performance in finding the true Pareto-optimal front, especially on some benchmark functions with many local Pareto-optimal fronts.     

Ensemble cluster pruning via convex-concave programming

Ensemble learning is the process of aggregating the decisions of different learners/models. Fundamentally, the performance of the ensemble relies on the degree of accuracy in individual learner predictions and the degree of diversity among the learners. The trade-off between accuracy and diversity within the ensemble needs to be optimized to provide the best grouping of learners as it relates to their performance. In this optimization theory article, we propose a novel ensemble selection algorithm which, focusing specifically on clustering problems, selects the optimal subset of the ensemble that has both accurate and diverse models. Those ensemble selection algorithms work for a given number of the best learners within the subset prior to their selection. The cardinality of a subset of the ensemble changes the prediction accuracy. The proposed algorithm in this study determines both the number of best learners and also the best ones. We compared our prediction results to recent ensemble clustering selection algorithms by the number of cardinalities and best predictions, finding better and approximated results to the optimum solutions.     

Multi-objective evolutionary algorithms based on the summation of normalized objectives and diversified selection

Most multi-objective evolutionary algorithms (MOEAs) use the concept of dominance in the search process to select the top solutions as parents in an elitist manner. However, as MOEAs are probabilistic search methods, some useful information may be wasted, if the dominated solutions are completely disregarded. In addition, the diversity may be lost during the early stages of the search process leading to a locally optimal or partial Pareto-front. Beside this, the non-domination sorting process is complex and time consuming. To overcome these problems, this paper proposes multi-objective evolutionary algorithms based on Summation of normalized objective values and diversified selection (SNOV-DS). The performance of this algorithm is tested on a set of benchmark problems using both multi-objective evolutionary programming (MOEP) and multi-objective differential evolution (MODE). With the proposed method, the performance metric has improved significantly and the speed of the parent selection process has also increased when compared with the non-domination sorting. In addition, the proposed algorithm also outperforms ten other algorithms.     

一种求解多目标优化问题的进化算法混合框架

为了提高多目标优化算法的求解性能,提出一种启发式的基于种群的全局搜索与局部搜索相结合的多目标进化算法混合框架.该框架采用模块化、系统化的设计思想,不同模块可以采用不同策略构成不同的算法.采用经典的改进非支配排序遗传算法(NSGA-II)和基于分解的多目标进化算法(MOEA/D)作为进化算法的模块算法来验证所提混合框架的有效性.数值实验表明,所提混合框架具有良好性能,可以兼顾算法求解的多样性和收敛性,有效提升现有多目标进化算法的求解性能.     

New mechanism for archive maintenance in PSO-based multi-objective feature selection

In classification problems, a large number of features are typically used to describe the problem’s instances. However, not all of these features are useful for classification. Feature selection is usually an important pre-processing step to overcome the problem of “curse of dimensionality”. Feature selection aims to choose a small number of features to achieve similar or better classification performance than using all features. This paper presents a particle swarm Optimization (PSO)-based multi-objective feature selection approach to evolving a set of non-dominated feature subsets which achieve high classification performance. The proposed algorithm uses local search techniques to improve a Pareto front and is compared with a pure multi-objective PSO algorithm, three well-known evolutionary multi-objective algorithms and a current state-of-the-art PSO-based multi-objective feature selection approach. Their performances are examined on 12 benchmark datasets. The experimental results show that in most cases, the proposed multi-objective algorithm generates better Pareto fronts than all other methods.     

Adapting exercise selection to performance,effort and self-esteem

ABSTRACT

Adapting to learner characteristics is essential when selecting exercises for learners in an intelligent tutoring system. This paper investigates how humans adapt next exercise selection (in particular difficulty level) to learner personality, invested mental effort, and performance to inspire an adaptive exercise selection algorithm. First, the paper describes the investigations to produce validated materials for the main studies, namely the creation and validation of self-esteem personality stories, mental effort statements, and mathematical exercises with varying levels of difficulty. Next, through empirical studies, we investigate the impact on exercise selection of learner's self-esteem (low versus high self-esteem) and effort (minimal, little, moderate, much, and all possible effort). Three studies investigate this for learners who had different performances on a previous exercise: just passing, just failing, and performed well. Participants considered a fictional learner with a certain performance, self-esteem and effort, and selected the difficulty level of the next mathematical exercise. We found that self-esteem, mental effort, and performance all impacted the difficulty level of the exercises selected for learners. Finally, using the results from the studies, we propose an algorithm that selects exercises with varying difficulty levels adapted to learner characteristics.     

一种维持种群多样性的多目标差分演化算法

差分演化算法是一种简单而有效的全局优化算法。本文将差分演化算法用于求解多目标优化问题,给出了一种维持种群多样性的多目标差分演化算法。该算法采用正交设计法初始化种群,改进差分演化算子,从而有利于维持种群多样性,提高演化算法的搜索性能。初步实验表明,新算法能有效地求解多目标优化问题。     

一种基于进化知识融合的多目标人工蜂群算法

提出一种基于进化知识融合的多目标人工蜂群算法.首先,采用精英群体知识和种群自身进化知识混合引导引领蜂进化,保持种群的多样性和优异性;然后,将一种融合个体支配关系和种群分布关系的方法引入跟随蜂的概率选择中,合理选择个体进行深度开发以改善算法收敛性能和分布性能;最后,提出一种更为严格的外部档案维护策略以降低外部档案维护成本,提高解集的分布性能.通过求解标准测试函数,并与其他3种多目标优化算法进行比较,仿真结果表明所提出算法具有良好的收敛性能和分布性能,且解集的覆盖范围更广.     

Learner-controlled selection of tasks with different surface and structural features: Effects on transfer and efficiency

Surface task features are more salient than structural task features and thus easier to recognize for novices. It is predicted that the more salient the task features the better learners can choose personally relevant and varied tasks, which enhances learning transfer. To investigate this prediction, a 2 × 2 factorial experiment with 72 participants studied the effects of control over tasks that differ in their surface features (learner, program) and in their structural features (learner, program). Learner control over the selection of tasks with salient surface features enables learners to select personally relevant and varied tasks. This is believed to yield higher effectiveness (i.e., higher near and far transfer test performance) as well as higher efficiency (i.e., higher transfer test performance combined with lower associated mental effort). Learner control over the selection of tasks with non-salient structural features does not enable learners to select personally relevant and varied tasks and is therefore not expected to yield beneficial effects on learning. The results show positive effects of learner control over the selection of tasks with salient surface features for efficiency on the far transfer test but not for effectiveness. Theoretical and practical implications are discussed.     

Genetic diversity as an objective in multi-objective evolutionary algorithms

A key feature of an efficient and reliable multi-objective evolutionary algorithm is the ability to maintain genetic diversity within a population of solutions. In this paper, we present a new diversity-preserving mechanism, the Genetic Diversity Evaluation Method (GeDEM), which considers a distance-based measure of genetic diversity as a real objective in fitness assignment. This provides a dual selection pressure towards the exploitation of current non-dominated solutions and the exploration of the search space. We also introduce a new multi-objective evolutionary algorithm, the Genetic Diversity Evolutionary Algorithm (GDEA), strictly designed around GeDEM and then we compare it with other state-of-the-art algorithms on a well-established suite of test problems. Experimental results clearly indicate that the performance of GDEA is top-level.     

A multi-objective artificial physics optimization algorithm based on ranks of individuals

This paper proposes a multi-objective artificial physics optimization algorithm based on individuals’ ranks. Using a Pareto sorting based technique and incorporating the concept of neighborhood crowding degree, evolutionary individuals in the search space are evaluated at first. Then each individual is assigned a unique serial number in terms of its performance, which affects the mass of the individual. Thereby, the population evolves towards the direction of the Pareto-optimal front. Synchronously, the presented approach has good diversity, such that the population is spread evenly on the Pareto front. Results of simulation on a number of difficult test problems show that the proposed algorithm, with less evolutionary generations, is able to find a better spread of solutions and better convergence near the true Pareto-optimal front compared to classical multi-objective evolutionary algorithms (NSGA, SPEA, MOPSO) and to simple multi-objective artificial physics optimization algorithm.     

一种双链结构的多目标进化算法DCMOEA

提出一种双链结构的多目标进化算法(DCMOEA).该算法采用双链结构表示个体,执行过程中无需设置外部归档集合,并采用ε支配策略保持解群的多样性.DCMOEA与MOEA/D、NSGA-II、SPEA2和PAES一同在4个2-目标ZDT函数和4个3-目标DTLZ问题上进行实验,并从算法所获解集的收敛性、分布均匀性和宽广性3个方面进行比较,仿真实验结果表明了DCMOEA的综合性能最好,是一种颇具竞争力的多目标进化算法.     

一种基于多样性度量的多目标化进化算法

针对如何通过附加的方法对多目标化问题进行理论分析,提出并证明了选择附加函数的3个前提条件.提出一种多目标化进化算法,根据种群中个体的多样性度量进行多目标化,并采用改进的非劣分类遗传算法对构造所得的多目标优化问题进行多目标优化.在静态和动态两种环境下进行算法性能验证,结果表明,在种群多样性保持、处理欺骗问题、动态环境下的适应能力等方面,所提算法明显优于其他同类算法.     

Feature-subspace aggregating: ensembles for stable and unstable learners

This paper introduces a new ensemble approach, Feature-Subspace Aggregating (Feating), which builds local models instead of global models. Feating is a generic ensemble approach that can enhance the predictive performance of both stable and unstable learners. In contrast, most existing ensemble approaches can improve the predictive performance of unstable learners only. Our analysis shows that the new approach reduces the execution time to generate a model in an ensemble through an increased level of localisation in Feating. Our empirical evaluation shows that Feating performs significantly better than Boosting, Random Subspace and Bagging in terms of predictive accuracy, when a stable learner SVM is used as the base learner. The speed up achieved by Feating makes feasible SVM ensembles that would otherwise be infeasible for large data sets. When SVM is the preferred base learner, we show that Feating SVM performs better than Boosting decision trees and Random Forests. We further demonstrate that Feating also substantially reduces the error of another stable learner, k-nearest neighbour, and an unstable learner, decision tree.     

基于全局排序的高维多目标优化研究

目标数超过4的高维多目标优化是目前进化多目标优化领域求解难度最大的问题之一,现有的多目标进化算法求解该类问题时,存在收敛性和解集分布性上的缺陷,难以满足实际工程优化需求.提出一种基于全局排序的高维多目标进化算法GR-MODE,首先,采用一种新的全局排序策略增强选择压力,无需用户偏好及目标主次信息,且避免宽松Pareto支配在排序结果合理性与可信性上的损失;其次,采用Harmonic平均拥挤距离对个体进行全局密度估计,提高现有局部密度估计方法的精确性;最后,针对高维多目标复杂空间搜索需求,设计新的精英选择策略及适应度值评价函数.将该算法与国内外现有的5种高性能多目标进化算法在标准测试函数集DTLZ{1,2, 4,5}上进行对比实验,结果表明,该算法具有明显的性能优势,大幅提升了4~30维高维多目标优化的收敛性和分布性.     

Dynamic multi-objective evolutionary algorithms for single-objective optimization

This paper proposes a new method for handling the difficulty of multi-modality for the single-objective optimization problem (SOP). The method converts a SOP to an equivalent dynamic multi-objective optimization problem (DMOP). A new dynamic multi-objective evolutionary algorithm (DMOEA) is implemented to solve the DMOP. The DMOP has two objectives: the original objective and a niche-count objective. The second objective aims to maintain the population diversity for handling the multi-modality difficulty during the search process. Experimental results show that the performance of the proposed algorithm is significantly better than the state-of-the-art competitors on a set of benchmark problems and real world antenna array problems.     

基于多目标优化策略的在线学习资源推荐方法

目前在线学习资源推荐较多采用单目标转化方法,推荐过程中对学习者偏好考虑相对不足,影响学习资源推荐精度.针对上述问题,文中提出基于多目标优化策略的在线学习资源推荐模型(MOSRAM),在学习者规划时间内,以同时获得学习者对学习资源类型偏好度最大和难度水平适应度最佳为优化目标,设计具有向邻居均值学习能力和探索新区域能力的多目标粒子群优化算法(NEMOPSO),提出以MOSRAM为核心的在线学习资源推荐方法(NEMOPSO-RA).不同问题规模下融合经典多目标优化算法的推荐方法对比实验表明,NEMOPSO-RA可以有效提高在线学习资源的推荐精度和推荐性能.     

A many-objective evolutionary algorithm based on hyperplane projection and penalty distance selection

For many-objective optimization problems, due to the low selection pressure of the Pareto-dominance relation and the ineffectivity of diversity maintenance scheme in the environmental selection, the current Pareto-dominance based multi-objective evolutionary algorithms (MOEAs) fail to balance between convergence and diversity. This paper proposes a many-objective evolutionary algorithm based on hyperplane projection and penalty distance selection (we call it MaOEA-HP). Firstly, the normalization method is used to construct an unit hyperplane and the population is projected onto the unit hyperplane. Then, a harmonic average distance is applied to calculate the crowding density of the projected points on the unit hyperplane. Finally, the perpendicular distance from the individual to the hyperplane as convergence information is added into the diversity maintenance phase, and a penalty distance selection scheme is designed to balance between convergence and diversity of solutions. Compared with six state-of-the-art many-objective evolutionary algorithms, the experimental results on two well-known many-objective optimization test suites show that MaOEA-HP has more advantage than the other algorithms, could improve the convergence and ensure the uniform distribution.     

基于自适应聚合距离的多目标进化算法

随着目标数的增多,种群收敛性与分布性的冲突愈加激烈,传统的多目标进化算法的选择算子难以平衡种群的收敛性与分布性.对此,提出一种基于自适应聚合距离的多目标进化算法.首先,采用参考点支配关系替代原有的Pareto支配关系,以增加选择压力,加强收敛性;其次,提出自适应聚合距离,通过动态变化的惩罚参数来自适应调整收敛性与分布性的比例;最后,设计一种带有淘汰算子的方法以改进小生境选择策略,根据自适应聚合距离的大小进行选择和淘汰操作.为验证算法的可行性,将所提出算法在测试问题上与其他4种优秀的多目标进化算法进行比较,并应用于两个实际应用中,仿真结果表明,所提出算法的综合性能更优,能有效平衡种群的收敛性与分布性.