The AND/OR process model exploits both AND and OR parallelism in logic programs. One of the major issues of the model is enhancing the intelligence of backtracking. In traditional methods, the backtracking has been based on the data dependency graph only, and consequently all variables, consumed by a failed literal, are regarded as the cause of the failure. However, a failure occurs due to not all of the variables in the data dependency graph but some variales in it. If we can get the variales which are the real cause of the failure, we can perform more intelligent backtracking using the result of the cause analysis.
In this thesis, the cause analysis which finds out the origin variables of failure and intelligent backtracking based on this cause analysis are re-examined and simulated.
