2-6. Non-temporal Experiment

As native short-circuit method make thorough examination in all numbers, on the event of a shutdown, backtracking of the search tree happens and it takes O(N!) of calculation cost in the worst case. A method developed to restrain this backtracking is the re-experimental graph method. In re-experimental graph method, it succeeds determinative paths deadlocked in the short-circuit graph and transform them into an alternative paths without deadlock. Where let points on determinative paths of short-circuit graph be succession points, when if we can make this alternative paths returned by re-experiment method include all of succession points, re-experiment method realize the re-experimentation never retrograding from any stage of short-circuit graphs.

To deal with deadlock, it is necessary to know about deadlock. With regard to deadlock, we will make detailed analysis in the next chapter and look over in what modality the non-determinability related to the deadlocks expands, but in this chapter we proceed the argument admitting temporarily that a rule called "overlapping principle of deadlock" holds. At first, using the overlapping principle of deadlock, we construct re-experimental graph method under the policy that we derive alternative paths which were properly possible from both of deadlock paths including all of possible cases and deadlock map. This method comes to be known impossible by reason of costs in the mean time, and in acknowledgment of that, we propose non-temporal experimentation to solve this problem essentially and try to formalize it. Non-temporal experimentation derives alternative paths without deadlock including all of succession points from deadlock map including remained possibility.