Questions for Week 9 Evolutionary Computation III
This week’s lab introduces a new modelling program, NetLogo.
1.
a. Using the default parameters (including grass off), run the model 8
times. For each trial record whether the both sheep and wolves go extinct, or
only wolves go extinct, or neither go extinct within 1000 generations.
b. Can you find a set of parameters (without setting the grass on) in which the
sheep and wolf populations reliably do not go extinct. Explain your result.
2.
a. Reset the model to the default parameters (reload the model if
necessary). Turn on the grass and set the initial number of wolves to zero. Run
the model several times (up to 500 generations). Do either sheep or grass go
extinct in any trials? Compare your results to those of question 1 and explain the
reasons for any differences.
b. Run one trial for 1000 generations and export the plot to an excel
spreadsheet. Calculate the average number of sheep for the last 500
generations.
c. Look up the description of the model and describe what variables are
represented by the sliders for grass-delay, init-sheep, sheep metabolism and
sheep-reproduce. Run the model several times and see if it is stable for
different parameter values.
d. Change each of the parameters in 2c in turn, keeping all other parameters
the same as in 2b. For each modification, run one trial for 1000 generations,
export the plot to an excel spreadsheet, and calculate the average number of
sheep for the last 500 generations.
e. Collate your results from 2b. and 2d into a table. Explain the similarities
and differences in the simulations for different parameters.
3.
Reset the model to the default parameters. Turn the grass on.
a. Describe the behaviour of the model with the three organisms.
b. Test a variety of parameters and describe what sets of parameter settings
would give
(i) fluctuating populations
(ii) stable non-fluctuating populations (non-extinct)
(iii) populations in which either or both sheep and wolves go extinct.
4. Try out other models from the models library, and describe one of the better ones. Explain briefly what it is designed to show and what results are given in the model.