Abstract
An ambulance location system was created for the British Co lumbia Provincial Ambulance Service. It was applied in the Capital Regional District (an area of about 937 square miles). The system helps administrators locate ambulances to improve response to calls.
The system consists of three parts: a means to reduce data to manageable proportions, a P-Median model for the determina tion of rough locations, and a computer simulation model for fine tuning of locations. The system worked well. The method of data reduction proved both practical and efficient. The rough locations provided by the P-Median model provided an excellent starting solution. The simulation model was shown to be ac curate in comparison with actual Capital Regional District data. The model was used with confidence to do a number of useful experiments concerning optimal ambulance locations and system response to changes in demand levels.
The validated system was used to answer many "what if" ques tions such as those that follow:
Are vehicles located in the best manner?
If we move one unit to reduce complaints from area A, how will this affect service in area B?
If we get more units, should we locate them incremen tally or would it be better to relocate all bases?
How much can we hedge against expected future demands?
How should we space shifts?
How should we determine the number of crews per shift?
Get full access to this article
View all access options for this article.
