Difference between revisions of "ParSA Results"

From UConn PAN
Jump to navigation Jump to search
Line 1: Line 1:
 
== Solutions ==
 
== Solutions ==
[[Image:testprob-5-2008.png|thumb|The test interferogram created from three random surfaces]]
+
[[Image:testprob-5-2008.png|thumb|The test interferogram created from three random surfaces.]]
[[Image:bestsol-5-2008.png|thumb]]
+
[[Image:bestsol-5-2008.png|thumb|The best solution found amongst several simulated annealing runs.]]
[[Image:truth1.png]]
+
[[Image:truth1.png|thumb|The front diamond surface used to generate the interferogram in the test problem.]]
[[Image:truth2.png]]
+
[[Image:truth2.png|thumb|The back diamond surface used to generate the interferogram in the test problem.]]
[[Image:demas1.png]]
+
[[Image:demas1.png|thumb|The best solution found amongst several simulated annealing runs.]]
[[Image:demas2.png]]
+
[[Image:demas2.png|thumb|The best solution found amongst several simulated annealing runs.]]
 
== Performance ==
 
== Performance ==
[[Image:alpha5plot2.png]]
+
[[Image:alpha5plot2.png|thumb|Log of probability versus non-convergence versus log of run length for ''alpha''=0.5.]]
[[Image:alpha9plot2.png]]
+
[[Image:alpha9plot2.png|thumb|Log of probability versus non-convergence versus log of run length for ''alpha''=0.9.]]

Revision as of 06:06, 2 June 2008

Solutions

The test interferogram created from three random surfaces.
The best solution found amongst several simulated annealing runs.
The front diamond surface used to generate the interferogram in the test problem.
The back diamond surface used to generate the interferogram in the test problem.
The best solution found amongst several simulated annealing runs.
The best solution found amongst several simulated annealing runs.

Performance

Log of probability versus non-convergence versus log of run length for alpha=0.5.
Log of probability versus non-convergence versus log of run length for alpha=0.9.