neuron-notes.txt

three types of neurons:
-sensory input
  output: 0.0 to 1.0
-internal neurons
  output: tanh(∑(inputs)) ie -1.0 to 1.0
-action outputs
  output: tanh(∑(inputs)) ie -1.0 to 1.0
  -> prob of firing

weight(for all): -4.0 to 4.0

1.start with all types of neurons
2.take genome to make connections ie brain
3.cull all neurons with no output if sensory/internal
 & cull all neurons with no input if action/internal
 (careful with self directed neurons)

SENSORY INPUTS
slr = pheromone gradient left-right
sfd = pheromone gradient forward
sg = pheromone density
DONE age = age
DONE rng = random input
blr = blockage left-right
bfd = blockage forward
osc = oscillator (simple sine wave, could be timed with day/year/etc.)
plr = population gradient left-right
pdf = population gradient foward
pop = population density
lpf = population long-range forward
lbf = blockage long-range forward
DONE lmy = last movement y
DONE lmx = last movement x
DONE bdy = north/south border dist
DONE bdx = east/west border dist
bd = nearest border dist
gen = genetic similarity of fwd neighbor
DONE lx = east/west world location
DONE ly = north/south world location

ACTION OUTPUTS
lpd = set long-probe dist
kill = kill forward neighbor
osc = set oscillator period
sg = emit pheromone
res = set responsiveness
mfd = move forward
mrn = move random
mrv = move reverse
mrl = move left/right (+/-)
mx = move west/east (+/-)
my = move north/south (+/-)