Xianneng Li
Papers
11
Total Citations
118
H-Index
7
About
Xianneng Li is a computational intelligence researcher whose work sits at the intersection of evolutionary algorithms, estimation of distribution algorithms (EDAs), and reinforcement learning. Over the course of more than a decade, Li has made sustained contributions to the development of graph-based evolutionary frameworks, most notably through advancing Probabilistic Model Building Genetic Network Programming (PMBGNP), a novel approach that represents candidate solutions as directed graph networks and constructs probabilistic models from promising individuals to guide evolution more effectively. A defining theme in Li's research is the creative use of traditionally discarded information. His work on incorporating infeasible or "bad" individuals into probabilistic model building challenged conventional truncation selection practices and demonstrated measurable improvements in problem-solving performance. He further enriched these frameworks by integrating reinforcement learning for probabilistic model construction, extending the approach to continuous optimization domains, and addressing population diversity loss to combat premature convergence. His most-cited paper on graph-based EDAs with reinforcement learning has garnered 37 citations, reflecting its influence on the field. More recently, Li has explored knowledge transfer in evolutionary computation and applied these principles to practical robotics challenges, including multi-target robotic arm control. His body of work reflects a consistent drive to make evolutionary algorithms smarter, more adaptive, and more applicable to real-world engineering problems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10