Investigate how to "chain" solution finding and improving to implement continuous evolution.

[cyber-pc]

Continued Solution Flow:

• Previous epoch = 197, current epoch = 198
• DERIVE (run once) at epoch 198
  - Use source's (pubkey_197, nonce_197, miningSeed_197) of previous epoch 197
  - initializeANN() + mutate(0..999)
  - Cache result as DerivedANN
• CONTINUE (for each new nonce) at epoch 198 :
  - Clone DerivedANN
  - Generate new initValue from (pubkey_198, nonce_198, miningSeed_198) for epoch 198
  - mutate(1000..1999)
  - Return score

[Franziska-Mueller]

As I mentioned in the meeting, it would be good to find good/optimal parameters for number of mutations and number of ticks/iterations outside of the Qubic chain where we don't have the constraints. Knowing these optimal parameters gives us a concrete goal to work towards and it also avoids doing any optimizations that will not pay off in the end.

Maybe ask CfB if the Anna dev did some experiments regarding optimal parameters.