My research interests evolve over time. Summarized below are some research questions that attracted me the most during the past few years.
Scalable Evolutionary Optimization
Scalability is a key issue for almost all computer algorithms. However, most meta-heuristic algorithms, e.g., evolutionary algorithms, scale poorly with the number of decision variables involved in an optimization problem. Hence, we have been working to develop novel scalable evolutionary algorithms, which could be used to tackle large-scale optimization problems. Specifically, (unlike some more classical computational problem, e.g., sorting) we are interested in problems that are not separable (decomposable) and that are characterized with non-continuous, non-differentiable, or even black-box objective functions. The core idea of our efforts along this direction is to explore smart method to decompose a problem in to sub-problems, such that a good solution to the original problem could be obtained by solving the related sub-problems. In short, our research can be regarded as a counter-part of divide-and-conquer method for non-separable large-scale (black-box) problems.
Selected relevant papers:
- Z. Yang, K. Tang and X. Yao, "Large Scale Evolutionary Optimization Using Cooperative Coevolution," Information Sciences, 178(15): 2985-2999, August 2008.
- Z. Yang, K. Tang and X. Yao, "Scalability of Generalized Adaptive Differential Evolution for Large-Scale Continuous Optimization,"Soft Computing, 15(11): 2141-2155, November 2011.
Learning and Optimization in Uncertain Environments
We are living in a world of uncertainty. This fact is, more often than not, exciting to me, since uncertainty may means some possibilities that had never come into my mind before. Hence, I've been always attracted by research topics that are relevant to the handling of uncertainties. Such topics (in various background) include, but not limited to:
Imbalanced Data Classification, where the "cost" of making mistakes on rare data is unknown or uncertain.
Evolutionary Robust/Dynamic Optimization, where the objective function and constraints may subject to uncertainty or even change over time.
- P. Wang, M. Emmerich, R. Li, K. Tang, T. Baeck and X. Yao, "Convex Hull-Based Multi-objective Genetic Programming for Maximizing Receiver Operating Characteristic Performance," IEEE Transactions on Evolutionary Computation, in press (DOI: 10.1109/TEVC.2014.2305671).
- M. Lin, K. Tang and X. Yao, "Dynamic Sampling Approach to Training Neural Networks for Multiclass Imbalance Classification," IEEE Transactions on Neural Networks and Learning Systems, 24(4): 647-660, April 2013.
Incremental Learning (highly relevant to online learning and stream data mining), where the concept (e.g., the underlying distribution of the data) may change over time.
- H. Fu, B. Sendhoff, K. Tang and X. Yao, "Robust Optimization Over Time: Problem Difficulties and Benchmark Problems," IEEE Transactions on Evolutionary Computation, in press. (DOI: 10.1109/TEVC.2014.2377125)
- F. Peng, K. Tang, G. Chen and X. Yao, "Population-based Algorithm Portfolios for Numerical Optimization," IEEE Transactions on Evolutionary Computation, 14(5): 782-800, October 2010.
- K. Tang, M. Lin, F. L. Minku and X. Yao, "Selective Negative Correlation Learning Approach to Incremental Learning," Neurocomputing, 72(13-15): 2796-2805, August 2009.
Data Driven Meta-heuristic Search
Meta-heuristic search algorithms iteratively generate and tests candidate solutions to a problem, and hence can be viewed as a data generating process itself. The generated data consist quite useful information about the problem to solve and the interaction between algorithms and problems. Making the best use of such data is a methodology that has the potential to boost the performance of the state-of-the-art meta-heuristic search algorithms. We've carried out quite a lot research along this direction, particularly to develop novel evolutionary algorithms for complex problems (e.g., large-scale problems or problems subject to uncertainty) that cannot be satisfactorily tackled by conventional EAs.
Selected relevant papers:
- P. Yang, K. Tang and X. Lu, "Improving Estimation of Distribution Algorithm on Multi-modalProblems by Detecting Promising Areas," IEEE Transactions on Cybernetics, in press. (DOI: 10.1109/TCYB.2014.2352411)
- L. Li and K. Tang, "History-Based Topological Speciation for Multimodal Optimization," IEEE Transactions on Evolutionary Computation, 19(1): 136-150, February 2015.
- K. Tang, F. Peng, G. Chen and X. Yao, "Population-based Algorithm Portfolios with automated constituent algorithms selection," Information Sciences, 279: 94-104, September 2014.
Capacitated Arc Routing
Capacitated Arc Routing Problems (CARPs) is a hard combinatorial problem that has broad applications in logistic and transportation domains. We have developed a number of novel EAs to various versions of CARP. Most of them have achieved the best-known solutions on a large variety of benchmark problems.
Selected relevant papers:
- K. Tang, Y. Mei and X. Yao, "Memetic Algorithm with Extended Neighborhood Search for Capacitated Arc Routing Problems," IEEE Transactions on Evolutionary Computation, 13(5): 1151-1166, October 2009.
- Y. Mei, K. Tang and X. Yao, "Decomposition-Based Memetic Algorithm for Multiobjective Capacitated Arc Routing Problem," IEEE Transactions on Evolutionary Computation, 15(2): 151-165, April 2011.
- Y. Mei, K. Tang and X. Yao, "A Memetic Algorithm for Periodic Capacitated Arc Routing Problem," IEEE Transactions on Systems, Man, and Cybernetics: Part B, 41(6): 1654-1667, December 2011.
- J. Wang, K. Tang, J. Lozano and X. Yao, "Estimation of Distribution Algorithm with Stochastic Local Search for Uncertain Capacitated Arc Routing Problems," IEEE Transactions on Evolutionary Computation, accepted on April 13, 2015.