Creating Difficult Instances of the Post Correspondence Problem Presenter: Ling Zhao Department of Computing Science University of Alberta March 20, 2001 Authored by Richard J. Lorentz Revised on May 3, 2001
Motivation Post Correspondent Problem (PCP) poses unique difficulties because of the unbounded size of search space Post Correspondent Problem (PCP) poses unique difficulties because of the unbounded size of search space Theoretically, PCP is the classic undecidable problem Theoretically, PCP is the classic undecidable problem The bounded version of PCP is NP- complete The bounded version of PCP is NP- complete Even the instance with small size and small width may have very long solution Even the instance with small size and small width may have very long solution
Introduction Definition Definition Given an alphabet and a finite set of pairs of strings ( g i, h i ) over the alphabet , does there exist a sequence i 1 i 2... i n of selections such that the strings g i 1 g i 2... g i n and h i 1 h i 2... h i n formed by concatenating are identical? width: the size of longest string in the pairs size: the number of pairs length of the solution: the size of the sequence of selection For simplicity, we restrict the alphabet to be {0, 1}
Example size = 3 width = 3
Example Solved! This instance has the shortest solution with length 7
Hard Instances size = 3 width = 3 length = 75 2 shortest solutions size = 4 width = 3 length = shortest solution
Main Work in the Paper Solve PCP’s Solve PCP’s 1. How to identify that the instance has no solution? 2. How to find the solution quickly? 3. How to search the solution space efficiently and quickly? Generate the hard instances of PCP Generate the hard instances of PCP 1. How to find the instances with small size and small width but long shortest length of solution?
Identify instances with no solutions Prefix / postfix filters Prefix / postfix filters Length balance filter Length balance filter Element balance filter Element balance filter
Prefix / postfix filters No starting point! No ending point!
Length / element balance filters The configuration will be always in the top! The configuration will be always in the top and contain several 1’s while any pairs can not decrease the number of it
Algorithms Involved Hash function and cache scheme Hash function and cache scheme Iterative deepening Iterative deepening Tail recursion removal Tail recursion removal
Recognize hard instances Generate PCP instances Generate PCP instances Iterative deepening Iterative deepening Random algorithm Random algorithm Restart when the number of searched nodes exceeds a threshold Restart when the number of searched nodes exceeds a threshold Use some heuristic information Use some heuristic information e.g. discourage the repetitions of pairs
My Progress on this problem Identify the situation that the configuration can not be on the top or in the below Identify the situation that the configuration can not be on the top or in the below Identify that one postfix in the configuration can not lead to the solution Identify that one postfix in the configuration can not lead to the solution Bidirectional search Bidirectional search Other branch pruning methods Other branch pruning methods
Top mask and Bottom mask If you want to make the configuration shrink its size to 0 or in the top, you can only use pair 1 At the beginning, the configuration is in the bottom Only two possible configurations satisfying the conditions: 101 No concatenations of the down strings can provide such postfixes! It has a bottom mask!
Useless Postfix DepthSelectionConfigurationProperty 1110Top Top Top Top Top Top Top After depth 1, if we can prove that all configurations will contain the postfixes either 110 or 101, we can infer that this instance has no solution
Useless Postfix DepthSelectionConfigurationProperty 1110Top Top Top Top Top Top Top S > S | S S > S | S Notes: I made a mistake for the proof of this postfix. Though it probably has such postfix, I can not prove it. This instance can be proved of no solutions by using the exclusion method. Added by Ling Zhao on May 3
Conclusions Use the standard techniques and simple heuristics to generate interesting PCP instance Use the standard techniques and simple heuristics to generate interesting PCP instance Raise many instances with the length larger than 100 Raise many instances with the length larger than 100 Conjectures about the minimum length of solutions for some instances Conjectures about the minimum length of solutions for some instances Conjectures about the most difficult instances for the specific width and size. Conjectures about the most difficult instances for the specific width and size.