论文部分内容阅读
平行结构类问题是一类适于分布式求解的人工智能问题．已有的大多数求解方法均采用预期或目标来指导自底向上的问题求解．但这些预期或目标是以局部问题求解状态为基础的，指导性较弱．尽管有的方法（如改进的ＤＶＭＴ结构）允许高层了解，但未给出明确的求解算法．本文提出一种双向求解平行结构类问题的方法，首先根据全局问题求解状态生成预期，指导自底向上的求解，然后根据新产生的假设来验证和修改预期，并用新的预期重新指导求解．该方法不仅提高了预期的指导性，而且使问题求解更为灵活．
The parallel structural problem is a kind of artificial intelligence problem that is suitable for distributed solving. Most of the existing solution methods use the expectation or goal to guide the bottom-up problem solving. However, these expectations or targets are based on the state of local problem solving and are weakly directed. Although some methods (such as improved DVMT structure) allow high-level understanding, but did not give a clear solution to the algorithm. In this paper, we propose a method to solve the parallel structure problem bidirectionally. First, we generate the expectation according to the global problem solving state, guide the bottom-up solution, and then validate and modify the expectation according to the newly generated hypothesis and redirect the solution with the new expectation. This method not only improves the expected guidance, but also makes the solution of the problem more flexible.