CSC 400 Team Project: Microbus Sytem
CSC 400 Team Project: Microbus Sytem
The first goal is to create a virtual world in which to develop and test
the system. Using existing cities as models, create a test city with
a population of 10,000 - 100,000 people, and a road network modeled
realistically regarding road capacities, speed limits etc.
Ideally, you should do traffic lights as well, but this may be too
much complexity, so make it a secondary goal.
You can use any existing software/data you can get your hands on.
Each inhabitant has a home address, and a large percentage
(50% or more) have a workplace and a work schedule.
In addition, each inhabitant has "random" errands that need to be run
daily (shopping, visits, recreation).
You will need a "character generator" to produce each inhabitant's
individual parameters.
A "level generator" for producing the city might be reasonable as well.
The dynamics of the city should be viewable with a "Sim-city" type
graphic interface. At a minimum, you need to show roads and
vehicles. Fancy 3-D graphics are not not needed
Once you have the simulator, investigate the feasibility and efficiency of
the Easy-Route proposal compared to individual cars and traditional
fixed-route busses. Important metrics are passenger transit time
(compared e.g. to time in individual cars)
and energy efficiency.
Important variables are the number and capacities of the busses.
For the purposes of this project, assume that the Microbus protocol
consists of the user calling the system at the time s/he needs a ride.
and that destination and current location are available at the time
of the call. You can consider modified versions if needed.
This is essentially a networking problem. The dynamic route planning
problem is difficult, and probably NP-hard in any non-trivial formulation,
so don't underestimate the importance of your scheduling algorithm.
Do some research on the subject before writing scheduling code.
Running the simulation could be time consuming as well.
This could be an excuse to make use of the department's computational
resources (come see me when you need them).
It might make sense to have a set of cities of various sizes
(e.g. from 100 on up) to investigate the scalability issues.
You should, of course also look into past work on problems of
this sort. There was a research project at MIT quite a few years ago
for instance. It has some significant limitations, but is worth looking at.