Problem: Emergency Medical Response

The Emergency Service Coordinator (ESC) for a county is interested in locating the countys three
ambulances to best maximize the number of residents that can be reached within 8 minutes of an
emergency call. The county is divided into 6 zones and the average time required to travel from one zone
to the next under semi-perfect conditions is summarized in the following Table 1.

Average Travel Times (min.)

Zones 1 2 3 4 5 6
1 1 8 12 14 10 16
2 8 1 6 18 16 16
3 12 18 1.5 12 6 4
4 16 14 4 1 16 12
5 18 16 10 4 2 2
6 16 18 4 12 2 2
Table 1: Average travel times from Zone i to Zone j in semi-perfect conditions.

The population in zones 1, 2, 3, 4, 5 and 6 are given in Table 2 below:

Zones Population
1 50,000
2 80,000
3 30,000
4 55,000
5 35,000
6 20,000
Total 270,000
Table 2: Population in each Zone

Goals of your model

1. Determine the locations for the three ambulances which would maximize the number of people who
can be reached within 8 minutes of a 911 call. Can we cover everyone? If not, then how many people are
left without coverage?
2. We now have only two ambulances since one has been set aside for an emergency call; where should
we put them to maximize the number of people who can be reached within the 8 minute window? Can we
cover everyone? If not, then how many people are left without coverage?
3. Two ambulances are now no longer available; where should the remaining ambulance be posted? Can
we cover everyone? If not, then how many people are left without coverage?
4. If a catastrophic event occurs in one location with many people from all zones involved, could the ESC
cover the situation? How do counties or cities design for those rare but catastrophic events?
5. In addition to the contests format, prepare a short 1-2 page non-technical memo outlining your
recommendations from your model and analysis finding for the ESC.
 Problem A

Emergency Medical Response

Team#4170

High School Mathematical Contest in Modeling

 Team #4170 Page 1 of 30

For office use only For office use only

T1 ________________ F1 ________________
T2 ________________ F2 ________________
T3 ________________ F3 ________________
T4 ________________ F4 ________________

2013

Summary Sheet

Team Control Number: 4170

Problem Chosen: A

Problem A: Emergency Medical Response

Summary
In order to best maximize the number of residents that can be reached as soon as
possible of an emergency call, our team strives to develop models to work out the best
solutions when simulating the real situation in the county. To simulate the realest situation,
we analyze the average travel time table and find the shortest time from one zone to
another.

Firstly, to determine the locations for 3 ambulances, we develop the first model not only
regardless of the cost times in a zone but also considering the cost of travel times in a
zone. When the cost times in the zone are ignored, ambulances located in Zone 2, 5 and
6 can cover the most residents, which is 300,000 people. In order to reduce the cost, we
locate the 3 ambulances in Zone 1, 2, and 5 or Zone1, 2, and 6, and all the zones can be
covered within 6 minutes. When consider travel times in a zone, 3 ambulances are
located in Zone 2, 4 and 5 or Zone 2, 5, and 6, the zones covered maximize. And there
are 275,000 people covered.

Secondly, after drawing the conclusion and giving out the answer of the first question, our
team further discuss the model how we determine the location of m ambulances in an
area which is divided into n zones. And we build a clear and detailed model which can be
used to almost every situation. It is also with regardless the travel time in a zone as well
as considering the travel time in a zone.
 Team #4170 Page 2 of 30

Then we substitute m, n into our new model. And we gain the best solutions of the second
and the third question using the second model. The simulation results validate that our
model is correct.

After that, we test our model in Shanghai. We choose 8 famous locations and analyze the
average time from one location to another. We also choose 3 locations as the starting
point. The result is that all the locations can be reached. This result proves that our model
is feasible in the real life.

At last, to solve the forth question, we turn the problem into a realistic example to analyze
it. At the time we verify our model, we use a real event as an example. We choose the
case of earthquake which happened in Wenchuan, China. We search a lot of information
on the Internet and get useful pictures and texts. After analyzing them, we make matrixes
and use our models to solve the problem. The simulation result is that most of the stricken
areas can be covered, but some roads are damaged so that several places cannot be
covered.
 Team #4170 Page 3 of 30

Contents
1. Problem Restatement .................................................................................................. 4
2. Assumptions and Justification................................................................................. 4
3. Analysis of the Average Time ................................................................................... 4
3.1 Variables Definition ...................................................................................... 5
3.2 Process for the Shortest Path ................................................................... 6
3.3 Conclusion ..................................................................................................... 6
4. Model Design ................................................................................................................. 8
4.1 Variables ......................................................................................................... 8
4.2 Model 1 ............................................................................................................ 9
4.2.1 Modeling .................................................................................................... 9
A. Locations regardless the travel time in a zone ............................ 9
B. Locations considering the travel time in a zone ........................ 11
4.2.2 Conclusion .............................................................................................. 13
4.3 Model 2 .......................................................................................................... 14
4.3.1 Modeling .................................................................................................. 14
A. Locations regardless of the travel time in a zone...................... 14
B. Locations considering the travel time in a zone ........................ 15
4.3.2 Conclusions ............................................................................................ 17
4.4 Model 3 .......................................................................................................... 17
4.4.1 Modeling .................................................................................................. 17
A. Locations regardless of the travel time in a zone...................... 18
B. Locations considering the travel time in a zone ........................ 19
4.4.2 Conclusion .............................................................................................. 20
4.5 Testing for the model ................................................................................ 21
5. The Solutions under Catastrophic Event ............................................................. 22
5.1 Variables ....................................................................................................... 22
5.2 Model ............................................................................................................. 22
5.3 Example ........................................................................................................ 23
6. Conclusion ................................................................................................................... 25
7. Appendix ...................................................................................................................... 26
a. solution_initial..................................................................................................... 26
b. solution_cover .................................................................................................... 27
c. solution_disaster................................................................................................ 28
d. The table of the data .......................................................................................... 29
 Team #4170 Page 4 of 30

1. Problem Restatement
What an ESC of a certain county does is to locate ambulances and dispatch them to a
particular place in the county which is divided into 6 zones.
According to the Average Travel Times Table, the time which it takes to travel from Zone i
to Zone j and the time that it takes from Zone j to Zone i are different. Maybe it is because
there are one-way roads or one-way traffic congestion. Besides, the time given by the
table which it takes to travel from one zone to another may be not the shortest. For
example, the time it takes from Zone 4 to Zone 5 is 16 minutes. But if the ambulance goes
to Zone 6 and then leave for Zone 5, it just takes 12 minutes. This is shorter than 16
minutes. The ESC has to maximize the number of people who can be reached within 8
minutes.
So as to maximum the people or zones to be covered, we make efforts to make and
improve our model so that we can simulate the situation more real.

2. Assumptions and Justification

Assumptions of Model 1, 2, 3:
1. Theres no accident on the way such as traffic jams, storms, etc.
2. The time that an ambulance travels from one spot to another is certain.
3. The population of each zone wont change.
4. Things such as machines of the ambulances run well.
5. There are hospitals in each zone.
6. Full fuel.

3. Analysis of the Average Time

It is easy to find a lot of ways from one zone to another in the picture, and their lengths are
different.
Tab. 1: The Original Average Time
Zone 1 2 3 4 5 7
1 1 8 12 14 10 16
2 8 1 6 18 16 16
3 12 18 1.5 12 6 4
4 16 14 4 1 16 12
5 18 16 10 4 2 2
6 16 18 4 12 2 2
 Team #4170 Page 5 of 30

The Original Picture of Routes with the Average Times

(When i<j, every red row shows the route from Zone i to Zone j, and every blue row shows
the route from Zone j to Zone i.)

3.1 Variables Definition

Variable Description
A matrix of average travel times from Zone i to Zone j in semi-perfect
A1
conditions
A matrix of the shortest travel times, regardless of the travel time in one zone,
A2
from Zone i to Zone j in semi-perfect conditions
A matrix with simplified information based on A2 which marks the number
A3 larger than 8 as 0 and the rest as 1 so that 1 means that the ambulance
can reach the place within 8 minutes
A matrix of the shortest travel times, considering the travel time in one zone,
A4
from Zone i to Zone j in semi-perfect conditions
A matrix with simplified information based on A4 which marks the number
A5 larger than 8 as 0 and the rest as 1 so that 1 means that the ambulance
can reach the place within 8 minutes
A matrix with simplified information based on A2 which marks the number
A6 larger than 6 as 0 and the rest as 1 so that 1 means that the ambulance
can reach the place within 6 minutes
A matrix with simplified information based on A2 which marks the number
A7 larger than 4 as 0 and the rest as 1 so that 1 means that the ambulance
can reach the place within 4 minutes

Pop A matrix which shows the population of each zone

 Team #4170 Page 6 of 30

1 8 12 14 10 16 1 8 12 14 10 12 1 1 0 0 0 0

8 1 6 18 16 16 8 1 6 18 12 10 1 1 1 0 0 0
12 18 1.5 12 6 4 12 18 1.5 10 6 4 0 0 1 0 1 1
A1 A2 A3
16 14 4 1 16 12 16 14 4 1 10 8 0 0 1 1 0 1
18 16 10 4 2 2 18 16 6 4 2 2 0 0 1 1 1 1

16 18 4 12 2 2 16 18 4 6 2 2 0 0 1 1 1 1

1 0 0 0 0 0 1 0 0 0 0 0
1 10 14.5 16 13 17

10 1 8.5 20 16.5 14.5
0 1 0 0 0 0
0 1 1 0 0 0
14.5 20.5 1.5 14.5 9.5 7.5 A 0 0 1 0 0 1 0 0 0 0 1 1
A4 5 0 0 1 A6
18 16 6.5 1 14.5 12.5 1 0 0 0 0 1 1 0 0

21 19 11.5 7 2 6 0 0 0 1 1 1 0 0 1 1 1 1

19 21 7.5 11 6 2 0 0 1 0 1 1 0 0 1 1 1 1

1 0 0 0 0 0

0 1 0 0 0 0
0 0 1 0 0 1
A7
0 0 1 1 0 0
0 0 0 1 1 1
0
0 1 0 1 1

Pop 50 80 30 55 35 20

3.2 Process for the Shortest Path

Critical Path Method:
In A1, we shorten the average travel time from one zone to another. For example, we use
Zone 2 to Zone 6 as series of data to get the shortest time. The number underrowd in the
table is the shortest time.
1 2 3 4 5 6
i=1 1
i=2 8 6 18 16 16
i=3 8 16 12 10
i=4 16 12 10

3.3 Conclusion
The shortest average time, regardless of the time in one zone, under semi-perfect
conditions from Zone i to j is showed in the following table.
Tab. 2:Shortest Average Travel Times Regardless the Travel Time in a Zone
Zone 1 2 3 4 5 6
1 1 8 12 14 10 12
2 8 1 6 18 12 10
3 12 18 1.5 10 6 4
4 16 14 4 1 10 8
 Team #4170 Page 7 of 30

5 18 16 6 4 2 2
6 16 18 4 6 2 2

The Picture of Routes with the Shortest Average Times Regardless of Travel Times
in a Zone
(When i<j, every red row shows the route from Zone i to Zone j, and every blue row shows
the route from Zone j to Zone i.)
The shortest time, considering the time in one zone, under semi-perfect conditions from
Zone i to j is showed in the following table.
Tab. 3: Shortest Average Travel Times Considering the Travel Time in a Zone
Zone 1 2 3 4 5 6
1 1 10 14.5 16 13 17
2 10 1 8.5 20 16.5 14.5
3 14.5 20.5 1.5 14.5 9.5 7.5
4 18 16 6.5 1 14.5 12.5
5 21 19 11.5 7 2 6
6 19 21 7.5 11 6 2
 Team #4170 Page 8 of 30

The Picture of Routes with Shortest Average Times Considering the Travel Time in
a Zone
(When i<j, every red row shows the route from Zone i to Zone j, and every blue row shows
the route from Zone j to Zone i.)

4. Model Design

4.1 Variables
In the optimal emergency medical response model, we define some variables as such:
Variable Name Description
k Index variable
n The number of zones
m The number of available ambulances

rk The number of people in Zone k

Zone k
vk

The location of the first ambulance

vx

The location of the second ambulance

vy

The location of the third ambulance

vz
 Team #4170 Page 9 of 30

vai The location where the i ambulance is located

The value of row x, column k in A3

s(vx , vk )

S(Vy,Vk) The value of row y, column k in A3

S(Vz,Vk) The value of row z, column k in A3
Sk The number of ambulances which can reach Zone k
yk Whether Zone k can be reached within 8 minutes
(Vx,Vk) The value of row x, column k in A5
(Vy,Vk) The value of row y, column k in A5
(Vz,Vk) The value of row z, column k in A5
zk Whether Zone k can be reached within 8 minutes
The variables in this table will be used in all of the following models.

4.2 Model 1
We have designed an optimal emergency medical response model for the determinations
of the locations of three ambulances.

4.2.1 Modeling
Under semi-perfect conditions and regardless of the travel time in one zone, we strive for
6
max yk rk
k 1

0, sk 0
yk
1, sk 1
s s (v , v ) s(v , v ) s(v , v )
k x k y k z k

vx 1, 2,3, 4,5, 6

v y 1, 2,3, 4,5, 6

s.t. vz 1, 2,3, 4,5, 6

s (v x , v k ) A3
s (v y , v k ) A3

s (v z , v k ) A3

vx v y vz
r Pop, k 1, 2,3, 4,5, 6
k

A. Locations regardless the travel time in a zone

Following the model, we substitute for every possible x, y, z. And we programme and gain

the following conclusion. (Program solution_initial is in the appendix; input Ak = A3 )

Table 4
The numbers of the ambulances which cover the The location of the
zone ambulances
No. 1 2 3 4 5 6
Team #4170 Page 10 of 30

1 2 2 2 1 1 1 1 2 5
2 2 2 2 1 1 1 1 2 6
3 1 1 2 1 1 2 1 3 4
4 1 1 2 1 2 2 1 3 5
5 1 1 2 1 2 2 1 3 6
6 1 1 2 2 1 2 1 4 5
7 1 1 2 2 1 2 1 4 6
8 1 1 2 2 2 2 1 5 6
9 1 1 3 1 1 2 2 3 4
10 1 1 3 1 2 2 2 3 5
11 1 1 3 1 2 2 2 3 6
12 1 1 3 2 1 2 2 4 5
13 1 1 3 2 1 2 2 4 6
14 1 1 3 2 2 2 2 5 6
(The first six columns of Table 4 mean the numbers of ambulances that cover this area.
The last three columns of D mean the location of 3 ambulances. There are 14 ways.)

a. The places where the ambulances locate that can cover the maximum
people regardless of the travel time in a zone
If an ambulance covers a whole zone, that ambulance covers everybody in this
zone. The solutions that have the maximum sum of the people who can be covered
by the three ambulances is the best solution.
We gain the best solution through Program solution_ initial. (Program solution_ initial is

in the appendix; input Ak = A3 ) So, the best solution is that when the 3 ambulances are

placed in Zone 2, Zone 5 and Zone 6, the number of the people who can be covered
maximize. And there are 300,000 people covered.
When theres a call, the 3 ambulances should go like what the following picture shows.

b. The places where the ambulances locate that cost least money or fuel
regardless of the travel time in a zone
If an ambulance costs least money and fuel, that ambulance spends least time
Team #4170 Page 11 of 30

traveling on the way.

Firstly, we use 6 minutes as a stand to further discuss A6 because the time of 6 minutes is
the longest time besides that of 8 minutes on the way. We programme through A6 and gain

the following conclusion. (Program solution_ initial is in the appendix; input Ak = A6 ) It is

that when we locate the 3 ambulances in Zone 1, Zone 2, Zone 5 or Zone1, Zone 2,
Zone 6, all the zones can be covered within 6 minutes. And the numbers of the people
covered are the same. There are 300,000 people covered.
Secondly, we use 4 minutes as a stand to further discuss A7 because the time of 4
minutes is the longest time besides that of 6 minutes on the way. We programme through
A7 and gain the following conclusion. (Program solution_ initial is in the appendix; input

Ak = A7 ) It is that ambulances cannot cover each zone within 4 minutes.

So, using this method, the best solution is to locate the 3 ambulances in Zone 1, Zone 2,
Zone 5 or Zone1, Zone 2, Zone 6. And there are 300,000 people covered.
When theres a call, the 3 ambulances should go like what the following pictures shows.

B. Locations considering the travel time in a zone

We get A4 in a situation that we consider the ambulance traveling in each zone. For
example, if an ambulance travels from Zone 1 to Zone 6, we should consider then sum of
time from Zone 1 to Zone 1, Zone 1 to Zone 6 and Zone 6 to Zone 6. Here is the new
expression.
Team #4170 Page 12 of 30

6
max zk rk
k 1

0 k 0
zk
1 k 1
(v , v ) (v , v ) (v , v )
k x k y k z k

vx 1, 2,3, 4,5, 6

v y 1, 2,3, 4,5, 6

s.t. vz 1, 2,3, 4,5, 6

(v x , v k ) A5
(v y , v k ) A5

(v z , v k ) A5

vx v y vz
r Pop, k 1, 2,3, 4,5, 6
k
We use 6 minutes as a stand and mark the number larger than 8 as 0 and the rest as 1
so that 1 means that the ambulance can reach the place within 8 minutes. After we
simplify the matrix, we get A5. We programme through A5 and gain the following

conclusion. (Program solution_ initial is in the appendix; input Ak = A5 ) It is that we

cannot cover every zone by 3 ambulances.

a. The places where the ambulances locate that can cover the maximum
people considering the travel time in a zone
We programme and gain the following conclusion. (Program solution_ initial is in the

appendix; input Ak = A5 )

The numbers of the ambulances which The location of The largest no.
cover the zone the of people
No. 1 2 3 4 5 6 ambulances covered
(thousand)
1 1 1 1 0 0 1 1 2 3 180
2 1 1 1 1 0 0 1 2 4 215
3 1 1 0 1 1 1 1 2 5 240
4 1 1 1 0 1 1 1 2 6 215
5 1 0 2 1 0 1 1 3 4 185
6 1 0 1 1 1 2 1 3 5 210
7 1 0 2 0 1 2 1 3 6 185
8 1 0 1 2 1 1 1 4 5 245
9 1 0 2 1 1 1 1 4 6 220
10 1 0 1 1 2 2 1 5 6 245
11 0 1 2 1 0 1 2 3 4 215
12 0 1 1 1 1 2 2 3 5 240
13 0 1 2 0 1 2 2 3 6 215
Team #4170 Page 13 of 30

14 0 1 1 2 1 1 2 4 5 275
15 0 1 2 1 1 1 2 4 6 250
16 0 1 1 1 2 2 2 5 6 275
17 0 0 2 2 1 2 3 4 5 245
18 0 0 3 1 1 2 3 4 6 220
19 0 0 2 1 2 3 3 5 6 245
20 0 0 2 2 2 2 4 5 6 280
We can get the best solution that when the 3 ambulances are located in Zone 4, Zone 5
and Zone 6, the people covered maximize. And there are 280,000 people covered.
When theres a call, the 3 ambulances should go like what the following picture shows.

b. The places where the ambulances locate that cover most zones
We programme and gain the following conclusion. (Program solution_ initial is in the

appendix; input Ak = A5 ) The best solution is that when the 3 ambulances are located in

Zone 2, Zone 4, Zone 5 or Zone 2, Zone 5, Zone 6, the zones covered maximize. And
there are 275,000 people covered.
When theres a call, the 3 ambulances should go like what the following pictures shows.

4.2.2 Conclusion
Team #4170 Page 14 of 30

We have turned a complex mathematical problem into a series of expression through a

model. And the problem has been worked out. At the same time, we have worked out 2
situations: regardless of the travel time in a zone or considering the travel time in a zone.
In the situations of Regardless of the travel time in a zone, we have further discussed the
two solutions to cost least money and fuel or to cover maximum people. Also, in the
situation of Considering the travel time in a zone, we have explored the two solutions to
cover most zones or people and give out different results.

4.3 Model 2
We have designed an optimal emergency medical response model for the determinations
of the locations of two ambulances.

4.3.1 Modeling
Under semi-perfect conditions and regardless of the travel time in a zone, we strive for
n
max yk rk
k 1

0, sk 0
yk
1, sk 1
m
sk s (vai vk )

i 1
s.t. v 1, 2,3, , n,i 1, 2,3, , m
ai
s (v ai , v k ) A3

va1 va2 vam

rk Pop, k 1, 2,3 , n

Its obvious that m equals 2 and n equals 6.

A. Locations regardless of the travel time in a zone

Following the model, we substitute for m, n and every possible x, y. And we programme
and gain the following conclusion. (Program solution_cover is in the appendix; input Ak =

A3 ; input Pop = Pop )

a. The places where the ambulances locate that can cover the maximum
people regardless of the travel time in a zone
We gain the best solution through Program solution_cover. (Program solution_cover is

in the appendix; input Ak = A3 ; input Pop = Pop ) It is that when we locate the two

ambulances in Zone 2, Zone 5 or Zone 2, Zone 6, the number of people covered

maximize. There are 300,000 people covered.
When theres a call, the 2 ambulances should go like what the following pictures shows.
Team #4170 Page 15 of 30

b. The places where the ambulances locate that cost least money or fuel
regardless of the travel time in a zone
If an ambulance costs least money and fuel, that ambulance spends least time
traveling on the way.
Firstly, we use 6 minutes as a stand because the time of 6 minutes is the longest time
besides that of 8 minutes on the way. We programme and gain the following conclusion.

(Program solution_cover is in the appendix; input Ak = A6 ; input Pop = Pop ) It is that

there exists no solution that cost less than that of 8 minutes.

B. Locations considering the travel time in a zone

We get A4 in a situation that we consider the ambulance traveling in each zone. For
example, if an ambulance travels from Zone 1 to Zone 6, we should consider then sum of
time from Zone 1 to Zone 1, Zone 1 to Zone 6 and Zone 6 to Zone 6. Here is the new
expression.
n
max zk rk
k 1

0, k 0
zk
1, k 1
m
k (vai vk )
i 1
s.t. v 1, 2,3, , ni 1, 2,3, , m
ai
(v ai , v k ) A5

va1 va2 vam

rk Popk 1, 2,3, , n
Its obvious that m equals 2 and n equals 6. We use 6 minutes as a stand and mark the
number larger than 8 as 0 and the rest as 1 so that 1 means that the ambulance can
reach the place within 8 minutes. After we simplify the matrix, we get A5. We programme
through A5 and gain the following conclusion. (Program solution_cover is in the appendix;

input Ak = A5 ; input Pop = Pop ) It is that we cannot cover every zone by 2 ambulances.
Team #4170 Page 16 of 30

a. The places where the two ambulances locate that can cover the
maximum people considering the travel time in a zone
We programme and gain the following conclusion. (Program solution_cover is in the

appendix; input Ak = A5 ; input Pop = Pop )

The numbers of the ambulances which cover The The largest no.
the zone location of people
of the covered
No. 1 2 3 4 5 6 ambulanc (thousand)
es
1 1 1 0 0 0 0 1 2 130
2 1 0 1 0 0 1 1 3 100
3 1 0 1 1 0 0 1 4 135
4 1 0 0 1 1 1 1 5 160
5 1 0 1 0 1 1 1 6 135
6 0 1 1 0 0 1 2 3 130
7 0 1 1 1 0 0 2 4 165
8 0 1 0 1 1 1 2 5 190
9 0 1 1 0 1 1 2 6 165
10 0 0 2 1 0 1 3 4 135
11 0 0 1 1 1 2 3 5 160
12 0 0 2 0 1 2 3 6 135
13 0 0 1 2 1 1 4 5 195
14 0 0 2 1 1 1 4 6 170
15 0 0 1 1 2 2 5 6 195
The best solution is that when the two ambulances are located in Zone 4, Zone 5 or Zone
5, Zone 6, the number of people covered maximize. There are 195,000 people covered.
When theres a call, the 2 ambulances should go like what the following pictures shows.

b. The places where the ambulances locate that cover most zones
We programme and gain the following conclusion. (Program solution_cover is in the
Team #4170 Page 17 of 30

appendix; input Ak = A5 ; input Pop = Pop )

The best solution is that when the two ambulances are located in Zone 4, Zone 5 or Zone
5, Zone 6, the number of people covered maximize. There are 195,000 people covered.
When theres a call, the 2 ambulances should go like what the following pictures shows.

4.3.2 Conclusions
Our team further discussed the model how we determine the location of m ambulances in
an area which is divided into n zones. At the same time, we have worked out 2 situations:
regardless of the travel time in a zone or considering the travel time in a zone. In the
situations of Regardless of the travel time in a zone, we have further discussed the two
solutions to cost least money and fuel or to cover maximum people. Also, in the situation
of Considering the travel time in a zone, we have explored the two solutions to cover
most zones or people and give out different results.

4.4 Model 3
We have designed an optimal emergency medical response model for the determinations
of the locations of an ambulance.

4.4.1 Modeling
Under semi-perfect conditions and regardless of the travel time in a zone, we strive for
Team #4170 Page 18 of 30

n
max yk rk
k 1

0, sk 0
yk
1, sk 1
m
sk s (vai vk )

i 1
s.t. v 1, 2,3, , n,i 1, 2,3, , m
ai
s (v ai , v k ) A3

va1 va2 vam

rk Pop, k 1, 2,3 , n

Its obvious that m equals 1 and n equals 6.

A. Locations regardless of the travel time in a zone

Following the model, we substitute for every possible x. And we programme and gain the

following conclusion. (Program solution_cover is in the appendix; input Ak = A3 ; input

Pop = Pop )

a. The places where the ambulance locate that can cover the maximum
people regardless of the travel time in a zone
We gain the best solution through Program solution_cover. (Program solution_cover is

in the appendix; input Ak = A3 ; input Pop = Pop ) It is that when we locate the ambulance

in Zone 2, the number of people covered maximize. There are 160,000 people covered.
When theres a call, the ambulance should go like what the following picture shows.

b. The places where the ambulances locate that cost least money or fuel
regardless of the travel time in a zone
We programme and gain the following conclusion. (Program solution_cover is in the
Team #4170 Page 19 of 30

appendix; input Ak = A3 ; input Pop = Pop ) It is that when we locate the ambulance in

Zone 5 or Zone 6, the number of people covered maximize. There are 140,000 people
covered.
When theres a call, the ambulance should go like what the following pictures shows.

B. Locations considering the travel time in a zone

We get A4 in a situation that we consider the ambulance traveling in each zone. For
example, if an ambulance travels from Zone 1 to Zone 6, we should consider then sum of
time from Zone 1 to Zone 1, Zone 1 to Zone 6 and Zone 6 to Zone 6. Its obvious that m
equals 1 and n equals 6. Here is the new expression.
n
max zk rk
k 1

0, k 0
zk
1, k 1
m
k (vai vk )
i 1
s.t. v 1, 2,3, , ni 1, 2,3, , m
ai
(v ai , v k ) A5

va1 va2 vam

rk Popk 1, 2,3, , n
We use 6 minutes as a stand and mark the number larger than 8 as 0 and the rest as 1
so that 1 means that the ambulance can reach the place within 8 minutes. After we
simplify the matrix, we get A5. We programme through A5 and gain the following

conclusion. (Program solution_cover is in the appendix; input Ak = A5 ; input Pop = Pop )

It is that we cannot cover every zone by 1 ambulance.

a. The places where the two ambulances locate that can cover the
maximum people considering the travel time in a zone
We programme and gain the following conclusion. (Program solution_cover is in the
Team #4170 Page 20 of 30

appendix; input Ak = A5 ; input Pop = Pop ) It is that when we locate the ambulance in

Zone 5, the number of people covered maximize. There are 110,000 people covered.
When theres a call, the ambulance should go like what the following picture shows.

b. The places where the ambulance locate that cover most zones
We programme and gain the following conclusion. (Program solution_cover is in the

appendix; input Ak = A5 ; input Pop = Pop ) It is that when we locate the ambulance in

Zone 5, the number of people covered maximize. There are 110,000 people covered.
When theres a call, the ambulance should go like what the following picture shows.

4.4.2 Conclusion
Just like model 2, we further discussed the model how we determine the location of m
ambulances in an area which is divided into n zones. At the same time, we have worked
out 2 situations: regardless of the travel time in a zone or considering the travel time in a
zone. In the situations of Regardless of the travel time in a zone, we have further
discussed the two solutions to cost least money and fuel or to cover maximum people.
Also, in the situation of Considering the travel time in a zone, we have explored the two
solutions to cover most zones or people and give out different results.
Team #4170 Page 21 of 30

4.5 Testing for the model

We prove that our model is correct via some examples.

4.5.1 Variables
Variables Description

A10 A matrix which describes the average times among the 8 famous location

A matrix with simplified information based on A11 which marks the number
A11
larger than 10 as 0 and the rest as 1 so that 1 means that the
ambulances can reach the place within 10 minutes

4.5.2 Example
Shanghai is an international metropolis. We choose 8 famous location from the center of
Shanghai and estimate the average time from one location to another. And we pick 3
suitable starting point where the 3 ambulances are located.
Dim Zone 1 as Xu Jia Hui
Dim Zone 2 as The Bund
Dim Zone 3 as West Shanghai Station
Dim Zone 4 as Jingan Temple
Dim Zone 5 as Hongkou stadium
Dim Zone 6 as Shanghai Circus Stage
Dim Zone 7 as Zhongshan Park
Dim Zone 8 as Wujiao Court

1 0 0 1 0 0 0 0
1 40 17 9 22 20 11 26
0 1 0 10 0 0 0
40 4 18 8 13 13 13 16
0 0 1 00 0 0 0
17 18 0.5 12 18 12 11 19

9 8 12 1 15 12 8 18 A 1 1 0 10 0 1 0
A10 11
0 0 0 01 1 0 1
22 13 18 15 2 9 21 9

20 13 12 12 9 1 16 10 0 0 0 01 1 0 1
11 13 11 8 21 16 2 21
0 0 0 10 0 1 0

26 16 19 18 9 10 21 4 1 1 0 1
0 0 0 0
Through programming we can draw the conclusion. (Program solution_cover is in the

appendix; input Ak = A11 , input Pop 1085 300 1130 349 792 707 690 1313 )

We get the result that when the 3 ambulances are located at Zone 3, Zone 4, Zone 5 or
Zone 3, Zone 4, Zone 8, all the zones can be covered. So, our model can be used in the
real situation and has a good feasibility.
Team #4170 Page 22 of 30

5. The Solutions under Catastrophic Event

Considering that a serious disaster happen, a lot of unexpected situations appear. The
roads may be destroyed so that ambulances would be delayed. The numbers of
population of each region are different, so there will be weights among regions in terms of
the regions importance. So there are a lot of things we cannot predict. Considering these
things, we simulate the situation and improve our model.

5.1 Variables
Variables Description
A matrix with each weight of each region in terms of the regions
M
importance

A8 The Average travel time from one city/county to another

A matrix with simplified information based on A8 which marks the number

A9 larger than 150 as 0 and the rest as 1 so that 1 means that the
ambulances can reach the place within 150 minutes

5.2 Model
Team #4170 Page 23 of 30

n
max mk yk rk
k 1

0, sk 0
yk
1, sk 1
m
sk s (vai vk )

i 1
s.t. v 1, 2,3, , n,i 1, 2,3, , m
ai
s (v ai , v k ) A8

va1 va2 vam

rk B, m k {m k , m k }, k 1, 2,3 , n

m k means the min weight which is given out. m k means the max weight which is given
out.

5.3 Example
To solve the forth question better, we turned the problem into a realistic example to
analyze it. And we chose the earthquake which happened in Wenchuan, China 5 years
ago.

Just like the process we solve Problem 1, 2, 3, we have to give out matrixes like A2 ,

Pop . So we chose 8 cities/counties randomly.

Team #4170 Page 24 of 30

We got the population of each city/county and the average time from one place to another
through the Internet. (The table of analyzing the data is in the appendix)We marked the
routes where the time is over 4 hours as the route which had been damaged and we
deleted these routes. And then we got the following matrix.
1 0 0 0 0 1 1 0
47 193 116 146

34 74 66 76 194 0 1 0 1 0 1 1 0
77 227 0 0 1 0 0 0 0 0

0 1 0 1 0 1 1 0
A8
74 52 94 103 172 A9
193 62 223 0 0 0 0 1 0 0 0

116 66 94 35 225 1 1 0 1 0 1 0 0
146 76 103 223 29 1 1 0 1 0 0 1 0

194 227 172 225 56 0 0 0 0 0 0 0 1

To sum up, we got the two matrix and information just like those of question 1, 2, 3.
After that, we worked out the average injured people of the 8 regions.
All the injured people . We made a program (Program
The average injured people
8

solution_disaster is in the appendix; input M k 1 1 2 1 1 2 2 1 ; input Ak A9 ; input

Pop 161 500 186 310 90 520 430 250 ) Finally, we get the result that when the
Team #4170 Page 25 of 30

ambulances are mostly in Place 1, 2, 4, 6, the people covered maximize and the
ambulances can reach the most serious stricken area.
But there exists disadvantages. First, there are less people in the most serious stricken
area. Second, maybe some places cannot be reached because of the damage of the
roads.
When theres a call, the ambulances should go like what the following pictures shows.

6. Conclusion
We build a simple model to solve the first problem. In the process, we gradually develop
the model and build our final model. Furthermore, we test our final model via the map of
Shanghai and the results are perfect. To solve the forth problem, we test our model again
via the map of Si Chuan Province and the simulation result is that most of the stricken
areas can be covered, but some roads are damaged so that several places cannot be
covered.
Team #4170 Page 26 of 30

7. Appendix

a. solution_initial
In Matrix C , each row means each solution. The first six columns means the numbers of
ambulances that cover this area. The last three columns of C means the location of the
3 ambulances. In Matrix D , each row means the sum number of the zones covered. In
Matrix P , each row means the sum number of the people covered by each ambulance.

And max_ P means the maximum of each element in P. Most importantly, every row of

each matrix means a certain solution.

Heres the code:
function [C,D, Cp, p, max_p]= solution_initial(Ak)
clc;
points = 6;
selected = 3;
RowNum = nchoosek(points,selected);
ColNum = points + selected;
C=zeros(RowNum,ColNum);
D=zeros(RowNum,1);
Cp = zeros(RowNum,points);
p=zeros(RowNum,1);
for i=1:points
m=1;
for j=1:4
for k=j+1:5
for l=k+1:6
C(m,i)= Ak(j,i)+Ak(k,i)+Ak(l,i);
C(m,7)=j;
C(m,8)=k;
C(m,9)=l;
m=m+1;
end
end
end
end
for i=1:RowNum
for u =1:points
if C(i,u)>0
D(i,1)=D(i,1)+1;
end
end
end
Cp(:,1)=C(:,1)*50;
Cp(:,2)=C(:,2)*80;
Team #4170 Page 27 of 30

Cp(:,3)=C(:,3)*30;
Cp(:,4)=C(:,4)*55;
Cp(:,5)=C(:,5)*35;
Cp(:,6)=C(:,6)*20;
for i=1:20
for j=1:6
p(i)=p(i)+Cp(i,j);
end
end
p;
max_p = max(p);

b. solution_cover
In Matrix C , each row means each solution. In Matrix D , each row means the sum
number of the zones covered. In Matrix S , each row means the location of the
ambulances. In Matrix P , each row means the sum number of the people covered by
each ambulance. And max_ P means the maximum of each element in P . Most
importantly, every row of each matrix means a certain solution.
Heres the code:
function [C,D, Cp,S, p,max_p] = solution_cover(points,locations,Ak,Pop)
clc;
X=[1:points];
S = combntns(X,locations);
len = size(S);
RowNum = nchoosek(points,locations);
C=zeros(RowNum,points);
D=zeros(RowNum,1);
Cp = zeros(RowNum,1);
p = zeros(RowNum,1);
for i = 1: len(1)
for j = 1:len(2)
C(i,:) = C(i,:) + Ak(S(i,j),:);
end
end
for i=1:RowNum
for k =1:points
if C(i,k)>0
D(i,1)=D(i,1)+1;
end
end
end
for i = 1 : points
Cp(:,i)=C(:,i)*Pop(i);
Team #4170 Page 28 of 30

end

for i=1:RowNum
for j=1:points
p(i)=p(i)+Cp(i,j);
end
end
max_p = max(p);

c. solution_disaster
In Matrix C , each row means each solution. In Matrix D , each row means the sum
number of the zones covered. In Matrix S , each row means the location of the
ambulances. In Matrix P , each row means the sum number of the people covered by
each ambulance. And max_ P means the maximum of each element in P . Most
importantly, every row of each matrix means a certain solution.
Heres the code:
function [C,D, Cp,S, p,max_p] = solution_disaster(points,locations,Ak,Pop,Mk)
clc;
X=[1:points];
S = combntns(X,locations);
len = size(S);
RowNum = nchoosek(points,locations);
C=zeros(RowNum,points);
D=zeros(RowNum,1);
Cp = zeros(RowNum,1);
p = zeros(RowNum,1);
for i = 1: len(1)
for j = 1:len(2)
C(i,:) = C(i,:) + Ak(S(i,j),:);
end
end
for i=1:RowNum
for k =1:points
if C(i,k)>0
D(i,1)=D(i,1)+1;
end
end
end
for i = 1 : points
Cp(:,i)=C(:,i)*Pop(i)*Mk(i);
end

for i=1:RowNum
Team #4170 Page 29 of 30

for j=1:points
p(i)=p(i)+Cp(i,j);
end
end
max_p = max(p);

d. The table of the data

 MEMORANDUM

From: the Ambulance Coordinating Interest Group

To: ESC

Date: 17/11/2013

Subject: Our recommendations from our model and analysis finding

As members of an ambulance coordinating interest group, we are interested in your

coordination of the medical ambulances.

Our team develops a simple model to place the three ambulances at the most efficient
locations. And we work out the best solutions to locate the 3 ambulances under
semi-perfect conditions regardless of the cost times in a zone but considering the cost of
travel times in a zone.

Then, we further discuss the model on how we determine the location of m ambulances in
an area which is divided into n zones, where m and n stand for not only 3 ambulances and
6 zones but any integer. So we substitute m, n for our new model.

Although the three ambulances might be enough to cover the whole county, but we do
recommend that you increase the number of ambulances because of the uncertain factors
that might affect different situations. Therefore, we developed this model to help you
decide how to put more of ambulances in the county.

Based on the development of technology, we suggest that you can improve the efficiency
of the engines in your ambulances to shorten the time of rescuing.

We also find that when natural disasters occur, the three ambulances will not be able to
save most people in time. If the only way between the two counties collapsed in such
disasters, more time will be wasted for making a detour or waiting for the cleaning of the
obstacles on the road. Therefore, we recommend that you should leave at least three
medical helicopters for saving people who live far away from the center of the county.
Earthwork cars are also important for cleaning the obstacles on the road which landslides.

Your consideration of this suggestion would be highly appreciated.

We are looking forward to seeing more people saved in disasters and difficult situations in
the county.