Guest Editorial Special Issue on Knowledge Extraction and Incorporation in Evolutionary Computation

TO ACQUIRE, understand, and reuse knowledge is one of the most important features of intelligent systems. Unfortunately, knowledge representation in humans and different machine systems could be very different, which makes it difficult to transfer knowledge between humans and machines, as well as between different machine systems. In order for the transferred knowledge to be reused, the knowledge must be represented in such a way that it is understandable to the user. Thus, the definition of interpretability of knowledge is user-dependent, and knowledge that is transparent to a machine system is not necessarily understandable to a human being, and vice-versa. If knowledge is to be transferred between humans and machines, fuzzy logic will play a key role [1]. Knowledge extraction and incorporation in evolutionary systems has received increasing interest in recent years. On the one hand, evolutionary algorithms have proved to be a powerful tool in extracting knowledge understandable to human beings in the form of symbolic or fuzzy rules from data. On the other hand, a priori or domain knowledge has shown to be very helpful, and even inevitable in many real-world applications to improve the efficiency of evolutionary algorithms. Knowledge can be incorporated in almost every element of evolutionary algorithms, such as representation, population initialization, crossover, and mutation, reproduction, fitness evaluation, and selection [2]. The target of this Special Issue is to put together the state-of-art and recent advances on knowledge extraction and incorporation in evolutionary computation. In response to our call for papers, 27 papers have been submitted. All submitted papers went through a peer-review procedure, and 14 papers have been selected to be included in the Special Issue based on the reviewers’ comments. The papers in the Special Issue can be roughly divided into two groups. The first group, including two papers and two correspondences, deals with knowledge extraction from data with the help of evolutionary algorithms. In the paper “A distributed evolutionary classifier for knowledge discovery in data mining,” by Tan et al., comprehensive symbolic rules are extracted using a distributed evolutionary algorithm, which is able to be implemented in different computers over the Internet. The paper “Agent-based evolutionary approach for interpretable rule-based knowledge extraction” by H. Wang et al. employs an agent-based evolutionary algorithm to extract interpretable fuzzy rules, where the tradeoff between accuracy and interpretability of the fuzzy system is addressed from a multiobjective optimization point of view. A fuzzy neural