Intro

Exploratory notebook related to Reactive Agents and Multiagents Systems. Includes toy implementations and visualization of the major introductory concepts for the topic. Content is strongly based and driven by Kadenze Online Course.

A software (or digital) agent is a computer program that acts autonomously (or semi-autonomously), and executes actions based on external and internal inputs.

Deliberative (or cognitive) agents maintain an internal state which influence their behavior.

Reactive agents don't maintain an internal state and act on external inputs in a reflexive manner, based on predefined behavioral sets.



In [ ]:

    
# Basic libraries import
import time
import numpy as np
import pdb
import sys
import os
from pathlib import Path
import seaborn as sns
from collections import namedtuple

# Project specific libraries
from scipy.spatial import distance
from PIL import Image, ImageDraw

# Plotting
import matplotlib
import matplotlib.pyplot as plt
from matplotlib import animation

sns.set_style("dark")
sns.set_context("paper")

%matplotlib notebook

%load_ext autoreload
%autoreload 2

Flocks

“Flocks, Herds, and Schools: A Distributed Behavioral Model” - Craig Reynolds 1987 SIGGRAPH paper. Here he referred to simulated flocks as boids.

Is about the behavior/movements simulation for a group of entities/units. Ideally there is no leader, and the flock behavior emerges from three simple rules:

Cohesion unit tendency to the average position of its neighbors
Alignment unit tendency to align itself with the average heading of its neighbors
Separation unit avoidance of collisions with its neighbors

Neighbors are only those entities of the group that a unit can perceive. This are defined based on a unit visibility.

Initially we will treat units as particles, so there is no need to take care of the rotation aspect of a unit.



In [ ]:

    
# import locally defined class
from Flock import Flock
CIRCLE_RADIUS = 2



In [ ]:

    
def draw_flow(flock: Flock, draw, circle_radius:int=CIRCLE_RADIUS, 
              visibility_radius=Flock.VISIBILITY_RADIUS):
    for unit in flock.units:
        x, y = unit.pos
        x_v, y_v = unit.vel
        # draw entity
        draw.ellipse([x-circle_radius, y-circle_radius, x+circle_radius, y+circle_radius], 
                     fill='black', outline='black')
        # draw visibility
        draw.ellipse([x-visibility_radius, y-visibility_radius, 
                      x+visibility_radius, y+visibility_radius], 
                     outline='red')
        # draw direction
        draw.line([x, y, (x+x_v*30), (y+y_v*30)], fill='blue')
        # draw velocity
        #draw.line([x, y, x+x_v, y+y_v], fill='blue')



In [ ]:

    
from Flock import Flock
img_size = 1000
flock = Flock(10, canvas_size=img_size/2)
img = Image.new('RGB', (img_size, img_size), (255, 255, 255))
fig, ax = plt.subplots(dpi=120, figsize=(5, 5))
im = ax.imshow(img)
plt.axis('off')

def animate(i, flock, img_size):
    img = Image.new('RGB', (img_size, img_size), (255, 255, 255))
    draw = ImageDraw.Draw(img)
    flock.update()
    draw_flow(flock, draw)
    im.set_data(img)

ani = animation.FuncAnimation(fig, animate, frames=500, interval=100, fargs=[flock, img_size])

Table of Contents

Intro

Flocks