8 - Bio Inspired AI

Question 1

Define a Boid

Answer 1

Reynold’s Boids simulate movement of a flock

• if too close to others, move away
• if too far, move toward
• if not aligned then align

Groups of birds have co-ordinated behaviours

Question 2

What is PSO - Particle Swarm Optimisation?

Answer 2

A bio inspired AI method inspired by the group foraging behaviour of bird flocks

• Comprised of several particles
• Each particle is a sim of bird and represents a point

Question 3

What defines particle movement trajectory?

Answer 3

A particle’s navigation trajectory is updated on the basis of its • Personal best: This is the particle’s best position (in terms of the evaluation/ objective function) so far in the iterative process.
• Neighbourhood best: This is the best position so far in the
neighbourhood of the particle

Question 4

What properties does each particle have?

Answer 4

position, value, velocity

Question 5

What is the PSO method?

Answer 5

1. init with n particles
2. for each, set p(best) - individual and n(best) - position of best particle
3. for each, calc velocity
4. update as per velocity
5. evaluate and update bests
6. if not end, go to step 3
7. return

Question 6

What sets particle velocity?

Answer 6

Start with init velocity
Velocity of particle in iteration t+1 is:
• xti : the particle’s current position during iteration t,
• vti : the particle’s current velocity during iteration t,
• xpbi : the particle’s personal best position, and
• xnbi : the current position of the best particle in the neighbourhood of the particle.

Question 7

What does PSO performance depend on?

Answer 7

num of particles
velocity updates
size and neighbourhood topology of the swarm