first thoughts about the project...

[HisMastersVoice]

hi miguel,

what about posting a small statement about the project ?

what could you tell us about the results you get back from the clients
are they better than you expected or are they just "in line" ?

or is the project into a too early stage to make statement right now ?

[michaelgarvie]

Certainly Your Mastery,
`Better than Human' Hall Of Fame (http://www.cogs.susx.ac.uk/users/mmg20/stats.php#hall) is a list of all the best results evolved so far, the names of the circuits were chosen by the owner of the island they came from...

These results are great. Evolved BIST solutions seem to require around half the overhead of conventional solutions!! This is scaling up as well. For eg. the Decod circuit required 83 gates for conventional full BIST while the evolved version only uses 52 (57 extra gates vs 26)!! However these are the smallest circuits of the industrial test suite, and they're all combinational.

We are currently attempting the evolution of a sequential circuit with BIST which takes a lot longer but has been increasing daily its BIST fitness by 0.01 so looks prômising. As soon as this is evolved I will write up a Journal paper on all these results. This paper will have the largest acknowledgements section ever...

For larger combinational circuits the circuit evaluation time goes up exponentially and the expected search space search time also increases exponentially. We still do not have enough participants to undertake such an experiment.

The aim of this project is to prove that evolved circuits can outperform conventional ones at any scale. It is also showing that a group of people can pool their resources to come up with stuff better than that done in industry, and then make it free for all. This is a great thing.

[HisMastersVoice]

hi miguel,

thanks for the answer.

i'll try to get some of my old teammates from other projects to join yours