Project Ideas
From enfascination
| Line 1: | Line 1: | ||
*Small world is a 'best of both worlds' between diameter of a random graph and clustering coefficient of a lattice. It takes very few random rewirings to give a lattice a random-graph's diameter. Is there a relationship between this and the particularly small value that mutation rate takes in GAs? Mutation rate is often conceptualized as a jump in the solution space, so it may not be hard to establish a connection. | *Small world is a 'best of both worlds' between diameter of a random graph and clustering coefficient of a lattice. It takes very few random rewirings to give a lattice a random-graph's diameter. Is there a relationship between this and the particularly small value that mutation rate takes in GAs? Mutation rate is often conceptualized as a jump in the solution space, so it may not be hard to establish a connection. | ||
| + | |||
| + | *My possible paths for immediate research (06/09): | ||
| + | **reimplement Kashtan and Alon | ||
| + | ***to look at decay on scaled up systems | ||
| + | ***to scale up systems, and perhaps look at hierarchy | ||
| + | **compare current evo with evo on independent legs, comparing mods | ||
| + | ***simple:calculate mods on arch004 bodies | ||
| + | **evolve topologies on 6-leg agents | ||
| + | **Get Paul's agent going | ||
| + | ***Get integration code up | ||
| + | ***get some walkers evolved | ||
Revision as of 21:35, 24 June 2009
- Small world is a 'best of both worlds' between diameter of a random graph and clustering coefficient of a lattice. It takes very few random rewirings to give a lattice a random-graph's diameter. Is there a relationship between this and the particularly small value that mutation rate takes in GAs? Mutation rate is often conceptualized as a jump in the solution space, so it may not be hard to establish a connection.
- My possible paths for immediate research (06/09):
- reimplement Kashtan and Alon
- to look at decay on scaled up systems
- to scale up systems, and perhaps look at hierarchy
- compare current evo with evo on independent legs, comparing mods
- simple:calculate mods on arch004 bodies
- evolve topologies on 6-leg agents
- Get Paul's agent going
- Get integration code up
- get some walkers evolved
- reimplement Kashtan and Alon