Wenjian Luo

Evolutionary clustering with differential evolution

Evolutionary clustering is a hot research topic that clusters the time-stamped data and it is essential to some important applications such as data streams clustering and social network analysis. An evolutionary clustering should accurately reflect the current data at any time step while simultaneously not deviate too drastically from the recent past. In this paper, the differential evolution (DE) is applied to deal with the evolutionary clustering problem. Comparing with the typical k-means, evolutionary clustering based on DE (deEC) could perform a global search in the solution space. Experimental results over synthetic and real-world data sets demonstrate that the deEC provides robust and adaptive solutions.

Differential evolution with a species-based repair strategy for constrained optimization

Evolutionary Algorithms (EAs) with gradient-based repair, which utilize the gradient information of the constraints set, have been proved to be effective. It is known that it would be time-consuming if all infeasible individuals are repaired. Therefore, so far the infeasible individuals to be repaired are randomly selected from the population and the strategy of choosing individuals to be repaired has not been studied yet. In this paper, the Species-based Repair Strategy (SRS) is proposed to select representative infeasible individuals instead of the random selection for gradient-based repair. The proposed SRS strategy has been applied to εDEag which repairs the random selected individuals using the gradient-based repair. The new algorithm is named SRS-εDEag. Experimental results show that SRS-εDEag outperforms εDEag in most benchmarks. Meanwhile, the number of repaired individuals is reduced markedly.

Combining multipopulation evolutionary algorithms with memory for dynamic optimization problems

Both multipopulation and memory are widely used approaches in the field of evolutionary dynamic optimization. It would be interesting to examine the effect of the combinations of multipopulation algorithms (MPAs) and memory schemes. However, since most of the existing memory schemes are proposed with single population algorithms, straightforwardly applying them to MPAs may cause problems. By addressing the possible problems, a new memory scheme is proposed for MPAs in this paper. In the experiments, several existing memory schemes and the newly proposed scheme are combined with a MPA, i.e. the Species-based Particle Swarm Optimizer (SPSO), and these combinations are tested on cyclic and acyclic problems. The experimental results indicate that 1) straightforwardly using the existing memory schemes sometimes degrades the performance of SPSO even on cyclic problems; 2) the newly proposed memory scheme is very competitive.

Immune genetic programming based on register-stack structure

Inspired by biological immune principles, a novel Immune Genetic Programming based on Register-Stack structure (rs-IGP) is proposed in this paper. In rs-IGP, an antigen represents a problem to be solved, and an antibody represents a candidate solution. A flexible and efficient antibody representation based on register-stack structure is designed for rs-IGP. Three populations are adopted in rs-IGP, i.e. the common population, the elitist population and the self set. The immune genetic operators are also developed, including clone operator, recombination operator, mutation operator, hypermutation operator, crossover operator and negative selection operator. The experimental results demonstrate that rs-IGP has better performance.

Intrusion detection oriented distributed negative selection algorithm

The negative selection algorithm proposed by Forrest et al. (1994) is a very significant change detection algorithm based on the generation process of T-Cells process in biological system. But when negative selection algorithm is used in distributed intrusion detection, the first problem that we meet is how to distribute the detectors in all detection workstations. To resolve this problem, this paper proposed a novel distributed negative selection algorithm based on the original negative selection algorithm. The core of this distributed negative selection algorithm is the distributing strategy. Two kinds of distributing strategies, random distributing strategy and greedy distributing strategy are given. Then we compared the performance of random distributing strategy and greedy distributing strategy. The experimental results show that: (1) distributed negative selection algorithm can avoid the problem of single point failure, when a part of detection workstations fails, the detection rate does not descend quickly; and (2) when some detection workstations fail, greedy distributing strategy can maintain better detection rate than random distributing strategy.

Clustering Time-Evolving Data Using an Efficient Differential Evolution

The previous evolutionary clustering methods for time-evolving data usually adopt the temporal smoothness framework, which controls the balance between temporal noise and true concept drift of clusters. They, however, have two major drawbacks: 1 assuming a fixed number of clusters over time; 2 the penalty term may reduce the accuracy of the clustering. In this paper, a Multimodal Evolutionary Clustering MEC based on Differential Evolution DE is presented to cope with these problems. With an existing chromosome representation of the ACDE, the MEC automatically determines the cluster number at each time step. Moreover, instead of adopting the temporal smoothness framework, we try to deal with the problem from view of the multimodal optimization. That is, the species-based DE SDE for multimodal optimization is adopted in the MEC. Thus the MEC is a hybrid of the ACDE and the SDE, and designed for time-evolving data clustering. Experimental evaluation demonstrates the MEC achieves good results.

Clustering spatial data by the neighbors intersection and the density difference

Clustering is a classical unsupervised learning task, which is aimed to divide a data set into several groups with similar objects. Clustering problem has been studied for many years, and many excellent clustering algorithms have been proposed. In this paper, we propose a novel clustering method based on density, which is simple but effective. The primary idea of the proposed method is given as follows. Firstly, the point with the largest local density in a cluster is considered as the cluster center. The local density of each point is estimated based on the distance (called radius) between the point and its k-th nearest neighbor. The point with a smaller radius indicates a larger local density. Secondly, the difference of the local densities between each two internal points should be small, while the difference between the density of a border point and the density of an internal point should be relatively large. Thirdly, if the intersection of k nearest neighbors of two points is small, they should be assigned to different clusters. The proposed algorithm has been compared with a typical clustering algorithm named FDPCluster, and the experimental results show that our algorithm has better clustering quality.

Clustering Evolutionary Data with an r-Dominance Based Multi-objective Evolutionary Algorithm

Clustering evolutionary data (or called evolutionary clustering) has received an enormous amount of attention in recent years. A recent framework (called temporal smoothness) considers that the clustering result should depend mainly on the current data while simultaneously not deviate too much from previous ones. In this paper, evolutionary data is clustered by a multi-objective evolutionary algorithm based on r-dominance, and the corresponding algorithm is named rEvoC. The rEvoC considers the previous clustering result (or historical data) as the reference point. We propose three strategies to define the reference point and to calculate the distance between a reference point and an individual. Based on the reference point and the r-dominance relation, the search could be guided into the region, in which a solution not only could cluster the current data well, but also does not shift two much from the previous one. Additionally, the rEvoC adopts one step k-means as a local search operator to accelerate the evolutionary search. Experimental results on two different data sets are given. The experimental results demonstrate that, the rEvoC achieves better performance than the corresponding static clustering algorithm and the evolutionary k-means algorithm.

Improved CFDP Algorithms Based on Shared Nearest Neighbors and Transitive Closure

A recently proposed clustering algorithm named Clustering by fast search and Find of Density Peaks (CFDP) can automatically identify the cluster centers without an iterative process. The key step in CFDP is searching for the nearest neighbor with higher density for each point. However, the CFDP algorithm may not be effective for cases in which there exist density fluctuations within a cluster or between two nearby clusters. In this study, two improved algorithms named CFDP-ED-TSNN1 and CFDP-ED-TSNN2 are presented, which adopt different ways to utilize the dissimilarity. Here, the dissimilarity is based on shared nearest neighbors and transitive closure. The experimental results on both several artificial datasets and a real-world dataset show that the improved algorithms are competitive.

Evolutionary design of polymorphic circuits with the improved evolutionary repair

In our previous work [1], the evolutionary repair technique has been introduced into the evolutionary design of the combinational logic circuits. In this paper, the evolutionary repair technique is improved, in which the number of the input vectors of the repair circuit is usually smaller than that of the corresponding incomplete circuit. The evolutionary algorithm with the improved evolutionary repair technique (i.e. erEDAII) is used to generate the polymorphic circuits. Experimental results demonstrate that some polymorphic circuits are evolved by erEDAII effectively. Especially, the polymorphic circuit with 8 inputs and 8 outputs could be evolved by the erEDAII.