Lecture 4

Swarm Intelligence

Why do animals swarm

• to forage better
• to migrate
• as a defense against predators
  • social insects have survived for millions of years
• swarms can achieve things that an individual cannot

Swarming examples

• Bird Flocking
• Boids model was proposed by Reynolds
  • bird - like (bird oids)
• Three rules while swarming
  • avoid collision with neighboring birds
  • match the velocity of neighboring birds
  • stay near neighboring birds

Swarming Characteristics

What is a swarm

• A loosely structured collection of interacting agents
  • Agents
    • individuals that belong to a group(but are not necessarily identical)
    • theyy contribute to and benefit from the group
    • they can recognize, communicate, and/or interact with each other
• The instinctive perception of swarms is a group of agents in motion - but that does not always have to be the case
• A swarm is better understood if thought of as agents exhibiting a collective behavior

Swarms examples in nature

• classic example
  • swarm of bees
• can be extended to other similar systems
  • ant colony
    • agents = ants
  • flock of birds
    • agents = birds
  • traffic
    • agents = cars
  • crowd
    • agents = humans
  • immune system
    • agents = cells and molecules

Homework

Topics for today

• softmax function
• introducton to Markov Decision Process

Softmax Function

• all values will be between 0 and 1
• I can apply them to prob where events are mutually exclusive and exhaustive

MDP

• most of times ( 90% ) I will move North
• this is what noisy env means

• I am also introducing some policies
  • policy means I have decided ki if I am in bottom 6 cells, I will move up and top 3 me I will move right
• with time, the reqard decreases
  • dollar today is btr than dollar tmrw
• here, all transitions have same time
  • chess ki tarah nahi hai ki kitna b time lo
  • but SMDP hota semi-mdp, usme time can also vary

To Read

• IBM Watson
• Softmax and sigmoidal
• read MDP