Evolutionary Computation in Constraint Satisfaction

摘要: Many difficult computational problems from different application areas can be seen as constraint satisfaction (CSPs). Therefore, plays an important role in both theoretical and applied computer science. Constraint deals essentially with finding a best practical solution under list of constraints priorities. methods, ranging complete systematic algorithms to stochastic incomplete ones, were designed solve CSPs. The methods are guaranteed the but usually perform great amount checks, being effective only for simple problems. Most these derived traditional backtracking scheme. Incomplete sometimes much faster; however, they not problem even if given unbounded time space. Because most real-world over-constrained do have exact solution, search is preferable deterministic methods. In this light, techniques based on meta-heuristics received considerable interest; among them, populationbased inspired by Darwinian evolution or collective behavior decentralized, self-organized systems, successfully used field satisfaction. This chapter presents some efficient evolutionary solving investigates development novel hybrid Satisfaction specific Evolutionary Computation paradigms. These approaches make use computation assisted inference algorithm. Comparative studies highlight differences between population-based performed Branch Bound