# Pushing cell agents
**URL:** https://discourse.processing.org/t/pushing-cell-agents/6077
**Category:** Processing.py
**Created:** [November 29, 2018, 7:10pm UTC](https://discourse.processing.org/t/pushing-cell-agents/6077 "2018-11-29T19:10:07Z")
**Posts on this page:** 6
**Page:** 1
### Author:  [@solub](https://discourse.processing.org/u/solub)
#### Post date: [November 29, 2018, 7:10pm UTC](https://discourse.processing.org/t/pushing-cell-agents/6077/1 "2018-11-29T19:10:08Z")
Hi,
I’m trying to make a Cell Division and Growth algorithm where cell agents are pushed right next to neighboring cell just after the division process. I’d like the final output to look like a circle packing.
However in my example sketch cells are pushed farther than their neighboring cell’s radii
How can I correct this behavior ?
```auto
def setup():
global collection
size(480, 360, P2D)
smooth(8)
collection = [Cell(PVector(width>>1, height>>1), 20)]
def draw():
background(255)
for c in collection:
c.growth()
c.interact()
c.display()
class Cell(object):
growthInterval = 10
divisionInterval = 100
def __init__ (self, pos, rad):
self.position = pos
self.radius = rad
self.time = 0
def interact(self):
for a in collection:
if a is not self:
d = PVector.dist(a.position, self.position)
if d < a.radius + self.radius:
#gap = (a.radius + self.radius) - d
diff = PVector.sub(self.position, a.position)
diff.normalize()
diff.div(d)
a.position.sub(diff)
def growth(self):
self.time += 1
if len(collection) < 100:
if self.time > 0 and self.time%Cell.divisionInterval == 0:
if random(1) > .5:
dir = PVector.random2D()
self.radius *= .5
child = Cell(self.position+dir, self.radius)
collection.append(child)
self.position.sub(dir)
if self.time%Cell.growthInterval == 0:
self.radius += .5
def display(self):
strokeWeight(3)
point(self.position.x, self.position.y)
noFill()
strokeWeight(1)
ellipse(self.position.x, self.position.y, self.radius, self.radius)
```
---
### Author:  [@solub](https://discourse.processing.org/u/solub)
#### Post date: [November 29, 2018, 8:08pm UTC](https://discourse.processing.org/t/pushing-cell-agents/6077/2 "2018-11-29T20:08:10Z")
So it seems I have to divide the radii by 2 (on the fourth line of the “interact” function) but it doesn’t make sense because I already did it on line 7 in the “growth” function.
If someone can find what’s going on.
---
### Author:  [@Sarah](https://discourse.processing.org/u/Sarah)
#### Post date: [November 29, 2018, 10:25pm UTC](https://discourse.processing.org/t/pushing-cell-agents/6077/3 "2018-11-29T22:25:52Z")
Here:
```auto
diff = PVector.sub(self.position, a.position)
diff.normalize()
diff.div(d)
a.position.sub(diff)
```
If you’re normalizing the vector, shouldn’t you multiply it by its radius instead? (Not divide it by the distance)
---
### Author:  [@solub](https://discourse.processing.org/u/solub)
#### Post date: [November 30, 2018, 11:57am UTC](https://discourse.processing.org/t/pushing-cell-agents/6077/4 "2018-11-30T11:57:37Z")
Hi Sarah,
I think that part (that I’m borrowing from Daniel Shiffman’s tutorial on boids) isn’t appropriate in this sketch and I shouldn’t be normalizing the the vectors’ distances in the first place. But you’re right in noting that I should _do something_ with the radii at some point (but don’t know what exactely).
Here’s what I came up with instead:
```auto
def interact(self):
for a in collection:
if a is not self:
#Distance between 2 vectors
d = PVector.dist(self.position, a.position)
#If that distance is smaller than the sum of the radii (+ strokeWeight)
if d < (a.radius*.5 + self.radius*.5 + 1):
#Compute gap distance
gap = (a.radius*.5 + self.radius*.5 + 1) - d
#Not sure what I'm doing here
diff = PVector.sub(self.position, a.position)
#Pushing the cell to the right distance (setting the magnitude to the sum of the radii)
a.position.sub(diff.setMag(gap))
```
Still, I don’t understand why I have to divide the radii by 2 again. It shouldn’t be that way.
---
### Author:  [@Sarah](https://discourse.processing.org/u/Sarah)
#### Post date: [November 30, 2018, 11:26pm UTC](https://discourse.processing.org/t/pushing-cell-agents/6077/5 "2018-11-30T23:26:19Z")
I think you might be mixing the radius with the diameter of the circle. The radius should be half of the circles diameter so instead of
`ellipse(self.position.x, self.position.y, self.radius, self.radius)`
it should be
`ellipse(self.position.x, self.position.y, self.radius*2, self.radius*2)`
---
### Author:  [@solub](https://discourse.processing.org/u/solub)
#### Post date: [December 1, 2018, 1:42pm UTC](https://discourse.processing.org/t/pushing-cell-agents/6077/6 "2018-12-01T13:42:01Z")
Of course… thank you Sarah.