Emanuel+Rosu+699+Homework

__Chapter 1__
1. Scenario: customer deciding queue to enter at store. Question: how can I get out of the store the fastest?

Model consists of customers, the number of items each has, cashiers; the main process would be checking out.

Scenario: store manager deciding how to operate queues for next hour. Question: how can I customers checked out most efficiently?

Model consists of customers, their variable number of items, employees available to cashier; again the main process would be checking out.

Scenario: store designer. Question: How can 100 customers per hour be checked out with the fewest employees. Model consists of the fixed number of customers, their variable number of items, a variable number of employees to see the effects of more or fewer being present; the main process, checking out, could be handled in multiple ways (1 big line, with available cashiers opening their checkout as available; always open lines with customers gravitating towards the shortest at the moment; always open lines, but with different item limits for different lines).

2. a) Need to represent trees, which can get varying degrees of sunlight (based on space for their canopy), and nutrients & water (based on a radius of roots around them). Because the trees and growth parameters are fixed it’s not a situation suited to an ABM.

b) Model would have to account for the lump savings sum, and the different rates available with each fund. This wouldn’t be a good situation to use an ABM because the rates for the individual funds don’t changed based on where the rest of the money goes; there is nothing dynamic or emergent.

c) Model would have to have variable lanes on a road and vehicle agents that navigate the road based on traffic and speed constraints. This would be a great use of an ABM as drivers react to different conditions.

d) Model would have to include whalers, a global variable tracking whales killed, and a whale population that is reactive to the actions of the whalers (how many whales they kill). An ABM could be used in this situation, but perhaps with a more focused question like “how does a whaling quota affect the whale population.” It would also be more useful if the whaler population also reacts to differing quotas and whale population sizes, as allowing more or less legal whaling would probably involve a change in the number of whalers.

e) Model would include a student body enrolling and available classes, with parameters such as class size and students having traits that decide enrollment priority. This would be an interesting application of an ABM, where you could see how students choose courses over the course of a degree to make sure they fulfill requirements, and one that could give insight into which classes need more availability.

f) Model would include a forest with trees and some type of representation of the trees being harvested. This seems a more mathematically oriented problem rather than one where an ABM would be useful.

g) Model would have to include banks that invest and customers who deposit and withdraw. It doesn’t seem a case particularly suited to an ABM because the customers aren’t exhibiting any learning or changing their behaviour--it would have to be “random”--and the banks would all be adhering to the government minimum percent, not adapting their behavior to what the customers do.

h) Model would include planes with varying carrying capacities and passengers, perhaps with varying parameters that they need to fulfill for flights (such as when they have to be in by). This would be suitable to an ABM as passengers would choose different flights based on what was available, allowing an observer to see which flights would fill up, if more flights fill up enough to make them worthwhile, etc.

i) Model would include planes and passengers; unlike the previous case, however, an ABM would not be suitable as the agents looked at are the planes themselves and not passengers, so there is little opportunity for an emergent situation to play out.

j) Model would include movie goers (with varying tastes) and the movies attended; this would not be a good use of an ABM as tastes won’t change in any non-random way based on anything that would be included in the model.

__Chapter 2__

Model:

Model w/ Edits:

3. Not always exactly 80 red patches, but always close (never seems to go over). This seems to happen when initial patch positions (of the 4) are close to each other, so I’m guessing the random 20 patches that then are turned red overlap.

__ Chapter 3 __

3. Simple ABM ODD:

Purpose: To observe the effect of human-induced habitat fragmentation on mating effectiveness in a certain animal species that can mate once a year.

Entities, State Variables, and Scales: In a very basic model, there would be 3 entities: animals, patches that represent the natural habitat of the bird (green patches), and patches that represent human activity & construction (brown patches). The animals would have variables for age and boolean variables that relay if the individual is ready to mate or now, based on time since last mated. Time steps would be take in months, and spatially both green and brown patches are simply considered either ‘natural’ or ‘human;’ the model won’t concern itself with specifics of what is in human or natural areas, except will assume that animals can never be in ‘human’ patches.

Processes: The main process is animal turtles searching for a mate, in basically a broad sweeping pattern (I’m imagining something similar to the mushroom hunter). Ideally something like the pheromones from the ants example shown in class could be implemented, so that when an individual is looking for a mate, it sends a signal within a radius around itself that signals other individuals within that radius to search that grid more minutely. The mating process occurs simply when two individuals wanting to mate are in adjacent green patches, and results in a new animal turtle being created and initialized at the location of one of its parents. Animal turtles die, or are removed from the field of play, after a pre-determined age is reached (each turtle’s age variable increments with time steps). The other process that would be occurring simultaneously would be the expansion of brown patches. At fixed intervals, brown patches expand by turning an adjacent green patch into a brown patch. Animals can now no longer place themselves on that patch. Throughout this, the observer would be able to track the how effectively this species is able to mate as their habitat is slowly fragmented; this could be achieved by a graphical representation tracking percent of animal turtles that are looking for a mate at each moment versus percent of the world that is ‘brown patches;’ the assumed outcome would be that the more brown patches there are, the higher the percent of animals that can’t find a mate at any given time is.

__ Chapter 4 __

Model with all changes implemented (changes commented in):

1. Prediction: q is 0.0, butterflies will move randomly; q is 1.0, butterflies will go directly to hilltops. Actual results match (although it’s worth noting that at 1.0 all the butterflies went up the high hill).

4. Movement is no longer so clearly directed towards hilltops, it takes agents more of a journey to get there.

__ 3D Model Attempt __

In collaboration with Craig and Collin; heavily based of the "hill climbers" code example in Netlogo.

Model:

Screenshot:



__Predator-Prey__

In collaboration with Collin:



__ Predator-Prey Revised __



Couple of BehaviorSpace graphs (5 runs w/ each pollution level):



__ Airport Models __





__ Airport Models (Revised) __





__ Airport Models - Final Slots Version (Hopefully) __



(check out the info tab inside the model for usage info and notes)

__ Turtles __

Link here.

__ Airport Modelling (Final) __

Netlogo:

Presentation: [|Link].