Grid Layout of Parking Population Modeling
- A set of Stores arranged in a line
- Each parking row is parallel to the line of stores
- Each row has the same number of parking spaces as there are stores
- The preference of each customer is assigned uniformly randomly
The parameters considered in the model are:
- The arrival rate of cars
- The departure rate of cars
- The strategy used by each car to choose the parking space depends on its proximity to the store of preference.
This graph shows a parking lot as a 10x10 grid where cars arrive with a Poisson probability distribution with mean of 10. There is no departure. The cars arrive and have a favorite column to park in which is linked to the preference of the store that the driver chooses. This is a Depth First Population model
DFP with no departure
Average Distance Summary:
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.00 2.00 5.00 4.95 8.00 10.00
The cars now arrive and fill up with a Breadth First approach
BFP with no departure
Average Distance Summary:
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.00 2.00 5.00 4.95 8.00 10.00
If the rate of departure is similar to the rate of arrival:
DFP with Same Departure Rate as Arrival - Poisson Arrival with mean 10
Average Distance Summary:
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.000 3.000 6.000 5.616 8.000 10.000
BFP with Same Departure Rate as Arrival - Poisson Arrival with mean 10
Average Distance Summary:
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.000 3.000 5.000 5.336 8.000 10.000
The models now consider a fixed rate of departure (1/10).
DFP width Fixed Departure rate (1/10) and Poisson Arrival with mean 10
Average Distance Summary:
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.00 3.00 6.00 5.72 8.00 10.00
BFP with Fixed Departure rate (1/10) and Poisson Arrival with mean 10
Average Distance Summary:
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.000 4.000 5.831 5.709 8.000 11.660
DFP with Fixed Departure rate (1/10) and Poisson Arrival with mean 100
Average Distance Summary:
Min. 1st Qu. Median Mean 3rd Qu. Max.
1.000 3.000 6.000 5.521 8.000 10.000
Observations
The results, while considering only a 10x10 grid, show that the average distance to the store of preference is greater when the drivers park using a Breadth First model than a Depth First model.
Under the assumptions of this model (uniform distribution of store preferences, distribution of arrivals...) the results show that when designing a parking lot, it appears that it would be more beneficial to lay the parking spots in a column style facing the stores rather than in a line.
Implementation
The model is implemented using the R language and is available here.
Last modified: Wed Jul 11 18:37:44 MDT 2007