Changes between Version 50 and Version 51 of Orio

Timestamp:

06/13/08 01:20:05 (15 years ago)

Author:

hartono

Comment:

--

Orio

@@ -196 +196 @@
 === Selecting Parameter Space Exploration Strategy ===
 
-Due to the huge search space of the parameter values, a proper search heuristic becomes a critical component of an empirical tuning system. Hence, in addition to an ''exhaustive search'' and a ''random search'', two effective and practical search heuristic strategies have been developed and integrated into the Orio’s search engine. These heuristics include the ''Nelder-Mead Simplex'' method and ''Simulated Annealing'' method. The exhaustive approach is selected as the default space exploration method of Orio. However, users can indicate the preferred search strategy in the tuning specifications, as exemplified in the ''search'' definition below.
+A conceptually straightforward approach to exploring the space of the parameter values is via an exhaustive search procedure. However, this exhaustive approach often becomes infeasible because the size of the search space can be exponentially large. Hence, a proper search heuristic becomes a critical component of an empirical tuning system. In addition to an ''exhaustive search'' and a ''random search'', two effective and practical search heuristic strategies have been developed and integrated into the Orio’s search engine. These heuristics include the ''Nelder-Mead Simplex'' method and ''Simulated Annealing'' method. The exhaustive approach is selected as the default space exploration method of Orio. However, Orio user can indicate his preferred search strategy in the tuning specifications, as exemplified in the ''search'' definition below.
 
 {{{
@@ -217 +217 @@
 To further improve the quality of the search result, each search heuristic is enhanced by applying a local search after the search completes. The local search compares the best performance with neighboring coordinates. If a better coordinate is discovered, the local search continues recursively until no further improvement is possible.  In the previous example, users can adjust the distance of the local search by modifying the value of the argument ''simplex_local_distance''. A local distance of two implies that the local search examines the performances of all neighbors within a distance of two.
 
+The following table lists all information about the search techniques implemented in the Orio's search engine.
+
+|| '''Search technique''' || '''Algorithm-specific arguments''' || '''Argument's default value''' || 
+|| ''Exhaustive'' || - || - ||
+|| ''Random'' || - || - ||
+|| ''Nelder-Mead simplex'' || `local_distance` (i.e. the maximum distance of neighboring coordinates considered by the local search) [[BR]] `reflection_coef` [[BR]] `expansion_coef` [[BR]] `contraction_coef` [[BR]] `shrinkage_coef` || - ||
+|| ''Simulated annealing'' || - || - ||
+
 == Developer Guide ==