计算机代写｜COM3524 Group-based Assignment – Applications of Cellular Automata

Before you start, please read the user documentation for the Python-based 1D and 2D Cellular Automata Simulator available on Blackboard or at http://pjworsley.github.io/capyle/. The 2D version of this custom-built software, effectively a Game of Life (GoL) simulator, will be your starting point the project. You need to understand its existing structure, then adapt it to solve the problem described below.

The aim of this assignment is to introduce you to the idea of how a Bioinspired Computational Method (Cellular Automata-based Simulation) can be applied to represent and investigate an important real world system: namely the spread of a forest fire.

After completing this project you should be able to work in a group to:

Note that though some of these skills are specific to this module, many of the above are useful for computer scientists and software engineers in general. You are not expected to have all these skills now but it is expected that you will learn a lot by interaction with the lecturer and demonstrators during the forthcoming lab sessions.

You are requested to submit a single A4 page Problem Analysis and Work Plan before the 2nd Forest Fire Computer Lab (15th November). This should include a list of all the tasks you have identified as being necessary to complete this assignment and a summary of what your collaborative approach will be (i.e. who does what and when – a Gantt Chart would be ideal). This plan will not receive a formal mark but is an important opportunity for feedback prior to completion and submission of your report. (email to P.Watton@sheffield.ac.uk before Monday 15th Nov).

Background: Modelling Wild Fires

The propagation of forest fires (wild fires) over particular terrains lends itself naturally to simulation with a CA. The array of cells represents the terrain over which the fire can spread, the state of each cell represents whether it is currently on fire (effectively “on” or “off”), and rules of varying complexity can be developed to model the fire spread.

In the simplest implementation, fires are initialised at specific positions on a grid of cells and each cell simply changes state to “on fire” if a neighbour is on fire. Additional complexity can be included by assigning a fuel resource parameter to each cell which is depleted over a number of generations, after which the fire goes out. This fuel parameter can be varied over the array of cells to represent different sorts of terrain. Depending on the timescale of the model, cells will either remain “off” indefinitely after the fuel level is depleted (though very long timescale models may include the concept of regrowth).

Additional levels of complexity could allow for any other factors that influence the way that the fire can spread. For instance, the neighbourhood rules can be adapted to introduce the concept of uncertainty by the use of stochastic-based decisions (e.g. the probability of propagation of the fire is based on the number of “on fire” neighbouring cells). In addition, neighbour probabilities can be adapted to model the effect of e.g. gradient and wind direction.

Statement of the problem

Figure 1: Schematic of region

A town in the USA is located in a region prone to fire fires. Specifically, town officials have become concerned about the possible fire risk posed to the town by a power generating plant that is situated approximately 30km to the north-east. A new waste incinerator – another potential source of ignition – has been proposed to be sited approximately 30km north-west of the town.

Between the town and these sites there are different types of terrain including:

A schematic of the region is shown in the figure below.