\addvspace {10\p@ }
\addvspace {10\p@ }
\addvspace {10\p@ }
\addvspace {10\p@ }
\contentsline {table}{\numberline {4.1}{\ignorespaces This table illustrates how HT4 can be mapped into the above definition of abduction.}}{32}
\contentsline {table}{\numberline {4.2}{\ignorespaces (Console et al 1991) versus (Menzies 1995)}}{34}
\contentsline {table}{\numberline {4.3}{\ignorespaces Hierarchy of Anomolies}}{38}
\contentsline {table}{\numberline {4.4}{\ignorespaces Sample experimental data from Smythe 87 model}}{52}
\addvspace {10\p@ }
\addvspace {10\p@ }
\contentsline {table}{\numberline {6.1}{\ignorespaces This table compares different versions of HT4.}}{87}
\addvspace {10\p@ }
\contentsline {table}{\numberline {7.1}{\ignorespaces This table compares the different stages of optimisation. Note that these results have been produced on a Sparc workstation running at 75MHz.}}{91}
\contentsline {table}{\numberline {7.2}{\ignorespaces This table compares different algorithmic optimisations made to the HT4 algorithm. Note that these results have been produced on an Intel Pentium PC, running at 133MHz, under Linux. Notice that the measurements for the transitive closure optimisation is not available, since this optimisation exceeded virtual memory on the PC.}}{96}
\contentsline {table}{\numberline {7.3}{\ignorespaces This table compares the different algorithmic optimisations made to the HT4 algorithm. Note that these results have been produced on a Sparc workstation, running at 50MHz, under SunOS. Notice that there is no difference in the mean explicables, and the mean percent reachable vertices.}}{97}
\contentsline {table}{\numberline {7.4}{\ignorespaces This table compares the differences between the randomisation studies and the normal C/C++ version of HT4. Note that these results have been produced on an Intel Pentium PC, running at 133MHz, under Linux. Notice how there is little difference in the mean percent explicable.}}{103}
\contentsline {table}{\numberline {7.5}{\ignorespaces This table compares the differences between the randomisation studies and the normal C/C++ version of HT4. Note that these results have been produced on a Sparc workstation, running at 50MHz, under SunOS. Notice how there is little difference in the mean percent explicable.}}{104}
\contentsline {table}{\numberline {7.6}{\ignorespaces This table compares the relative times spent in the forward sweep of the HT4 algorithm, before and after the refinement of the requirements. Note that these results have been produced on an Intel PC running at 133MHz, under Linux.}}{106}
\addvspace {10\p@ }
\addvspace {10\p@ }