A Hybrid Evolutionary Algorithm For Maximizing Satisfiability In Temporal Or Spatial Qualitative Constraints
10TH HELLENIC CONFERENCE ON ARTIFICIAL INTELLIGENCE (SETN 2018)(2018)
摘要
In this paper we tackle the MAX-QCN optimization problem, which consists in characterizing a consistent scenario that maximizes the satisfiability of a spatial or temporal qualitative constraint network (QCN). We propose an original hybrid evolutionary algorithm for solving the MAX-QCN problem, which we call EAMQ for short. This EAMQ method consists in randomly generating an initial population of consistent G-scenarios, and then realizing in an iterative manner an evolution of this population by generating new G-scenarios from crossover operations applied on the better individuals of the population at hand. Additionally, every time a new scenario is generated, an exploration of its neighborhood is realized in order to obtain a better scenario. Preliminary experiments conducted on QCNs of the Interval Algebra show the interest of our approach for solving the MAX-QCN problem.
更多查看译文
关键词
Spatio-temporal reasoning, qualitative constraints, optimization
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络