You are on page 1of 12

4. Consider the possibility of building a reservoir upstream of three demand sites along a river.

The net benefits derived from each use depend on the reliable amounts of water allocated to each
use. Letting xit be the allocation to use i in period t, the net benefits for each period t equal
1. 6x1t– x1t

2. 7x2t – 1.5 x2t

3. 8x3t – 0.5 x3t


Assume the average inflows to the reservoir in each of four seasons of the year equal 10, 2, 8, 12.
Assuming a reservoir capacity of 5, and dividing the release into integer increments of 2 (i.e., 2,
4, 6 and 8), using DP find the optimal operating policy. Assume the maximum release cannot
exceed 8, and the minimum release cannot be less than 2. Assume initial storage of 0 of the
reservoir. Solve the problem analytically (using EXCEL Spread sheet) and develop a computer
code to solve this specific problem.

USER 1
USER 3
Inflow

USER 2
GIVEN
→ A resorvior upstream of water users
→ Number of water users(Demand sites) =3
→ If water is released from the resorvior ; the net benefit gained fom user is:-
User-1=6X1t -X1t 2
User-2=7X2t -1.5X2t 2
User-3=8X3t -0.5X3t 2
→ Inflow to the resovior for four seasons of ayear are:-
Q1 = 10
Q2 = 2
Q3 = 8
Q4 = 12
→ Resorvior capacity,K=5
→ Release is done in integers of increment 2, (i.e. 2,4,6,8 )
→ initial storage of the resorvior, S1=0
REQUIRED
→ Optimal operating policy of the resorvior
SOLUTION
→ For all possible values of release (i.e.2,4,6,8),we have to work the net benefit
gained by optimaly allocating the releases.
then after,setteing the relationship between RELEASE and BENFIT the resorvior release policy
(i.e. how much amount of water should be released for different seasons can be fixed).
→ Now,lets dtermine the water allocation for the three users.
a) WATER ALLOCATION FOR THREE USERS (forwared recursion)
→ s1=Available water at stage 1
s2=Available water at stage 2
s3=Available water at stage 3
f1*(s1)=maximum return gained due to allocation of s1
f1*(s2)=maximum return gained due to allocation of s2
f1*(s3)=maximum return gained due to allocation of s3
→ Assuming allocation is done to users in integer of increment of 2 the feasible allocations are 0,2,4,6,8
Stage-1 2
0 ≤ x1 ≤ s1
0 ≤ x1 ≤ 8
0 0
2 8
6 8 2,4
4 8
6 0
0 0
2 8
8 4 8 8 2,4
6 0
8 -16

Stage-2
0 ≤ x2 ≤ s2
0 ≤ x2 ≤ 8
f2*(s2)=max (R2(x2)+f1*(s2-x2))
2
= max(f1*(s2-x2)+(7X2t -1.5X2t ))
.(1) .(2) .(3) .(4) .(5) .(6) .(7) .(8)
s2 x2 R2(x2) (s2-x2) f1*(s2-x2) (3)+(5) f2*(s2) x2*
0 0 0 0 0 0 0 0
0 0 2 8 8
2 8 0,2
2 8 0 0 8
0 0 4 8 8
4 2 8 2 8 16 16 2
4 4 0 0 4
0 0 6 8 8
2 8 4 8 16
6 16 2
4 4 2 8 12
6 -12 0 0 -12
0 0 8 8 8
2 8 6 8 16
8 4 4 4 8 12 16 2
6 -12 2 8 -4
8 -40 0 0 -40

Stage-3
0 ≤ x3 ≤ s3
0≤x3≤8
f3*(s3)=max (R3(x3)+f2*(s3-x3))
= max(f2*(s3-x3)+(8X3t -0.5X3t 2 ))

3
.(1) .(2) .(3) .(4) .(5) .(6) .(7) .(8)
s3 x3 R3(x3) (s3-x3) f2*(s3-x3) (3)+(5) f3*(s3) x3*
0 0 0 0 0 0 0 0
0 0 2 8 8
2 14 2
2 14 0 0 14
0 0 4 16 16
4 2 14 2 8 22 24 4
4 24 0 0 24
0 0 6 16 16
2 14 4 16 30
6 32 4
4 24 2 8 32
6 30 0 0 30
0 0 8 16 16
2 14 6 16 30
8 4 24 4 16 40 40 4
6 30 2 8 38
8 32 0 0 32

→ If Available water at stage-3 (i.e.release from the resorvior is s3 = 8), for optimal use from table stag
Allocation to user-3, X3 =4
s2* = s3-X3 = 8-4 = 4
→ From table stage-2
for s2=4
X2 *=2 and;
→ From table stage-1
X1 *=2
→ Therefore the water allocation to the three users shoul be;
ͮ→ User-1 = X1 = 2 units of water
ͮ→ User-2 = X2 = 2 units of water
ͮ User-3 = X3 = 4 units of water

→ Similarly if available water (reiease) is 6 units
water allocation to the three users shoul be;
ͮ→ User-1=X1 =0 units of water
ͮ→ User-2=X2 =2 units of water
ͮ User-3=X3 =4 units of water

→ if available water (reiease) is 4 units
ͮ→ User-1=X1 =0 units of water
ͮ→ User-2=X2 =0 units of water
ͮ→ User-3=X3 =4 units of water

4
→ if available water (reiease) is 2 units
ͮ→ User-1=X1 =0 units of water
ͮ→ User-2=X2 =0 units of water
ͮ→ User-3=X3 =2 units of water

→ Now we can set relationship between BENEFIT and RELEASE


RELEASE BENEFIT
0 0
2 14
4 24
6 32
8 40

b) RESORVIOR RELEASE POLICY (backwared recursion)

t=1 t=2 t=3 t=4


n=4 n=3 n=2 n=1
S1 = 0 S2 S3 S4
Q 1 =10 Q2= 2 Q3 = 8 Q4 = 12

St+1 = St + Q t - Rt
n
ft (St ) = Maximized net benefit
→ Stage-1
n= 1 S4
t=4 R4
Q 4 = 12
f4 1 (S4 ) = Max[B4 (R4 )]
0 ≤ R4 ≤ S4 + Q 4
S4 + Q 4 - R4 ≤ 5
→ For S4 = 0 and Q 4 = 12; R4 ≥ 7 and; Take R4 = 8
For S4 = 2 and Q 4 = 12; R4 ≥ 9 Take R4 = 8
Since we are ristericted the release to be maximum of 8,we should have to take R4 = 8.
this means that the extra unit of water should be considred as an over flow.
→ In a similar fashion the possible releases and their corresponding benefits are tabulated below.

5
1
S4 R4 B4 (R4 ) f4 (S4 ) R4
0 8 40
1 8 40
2 8 40
40 8
3 8 40
4 8 40
5 8 40

Stage-2
n= 2 S3
t=3 R3
Q3 = 8
f3 2 (S3 ) = Max[B3 (R3 ) + f4 1 (S3 +Q 3 -R3 )]
0 ≤ R3 ≤ S3 + Q3
S3 + Q3 - R3 ≤ 5
.(1) .(2) .(3) .(4) .(5) .(6) .(7) .(8)
1 2
S3 R3 B3(R3) S3+Q3-R3
f4 (S3+Q3-R3)
(3)+(5) f3 (S3 ) R3*
4 24 4 40 64
0 6 32 2 40 72 80 8
8 40 0 40 80
4 24 5 40 64
1 6 32 3 40 72 80 8
8 40 1 40 80
6 32 4 40 72
2 80 8
8 40 2 40 80
6 32 5 40 72
3 80 8
8 40 3 40 80
4 8 40 4 40 80 80 8
5 8 40 5 40 80 80 8

Stage-3
n= 3 S2
t=2 R2
Q2 = 2
f2 3 (S2 ) = Max[B2 (R2 ) + f3 2 (S2 +Q 2 -R2 )]
0 ≤ R2 ≤ S2+Q2
S2 +Q2 - R2 ≤ 5

6
.(1) .(2) .(3) .(4) .(5) .(6) .(7) .(8)
2
S2 R2 B2(R2) S2+Q2-R2 f3 (S2+Q2-R2) (3)+(5) f2 3 (S2 ) R2*
0 2 14 0 80 94 94 2
1 2 14 1 80 94 94 2
2 14 2 80 94
2 104 4
4 24 0 80 104
2 14 3 80 94
3 104 4
4 24 1 80 104
2 14 4 80 94
4 4 24 2 80 104 112 6
6 32 0 80 112
2 14 5 80 94
5 4 24 3 80 104 112 6
6 32 1 80 112
Stage-4
n= 4 S1 = 0
t =1 R1
Q1 = 10
f1 4 (S1 ) = Max[B1 (R1 ) + f2 3 (S1 +Q1 -R1 )]
0 ≤ R1 ≤ S1 +Q1
S1 + Q1 - R1 ≤ 5
.(1) .(2) .(3) .(4) .(5) .(6) .(7) .(8)
3 4
S1 R1 B1(R1) S1+Q1-R1 f2 (S1+Q1-R1) (3)+(5) f1 (S1 ) R1*
6 32 4 112 144
0 144 6,8
8 40 2 104 144

→ Therefore, from this table For S1 = 0 and Q1 = 10;


Realese ,R1 = 6 or 8
→ Tracing back and using resorvior continutiy equation ;
St+1 = St + Q t - Rt
→ S2=S1+Q1-R1
S2=10+0-6=4 → This gives R2*= 6 and/or
S2=10+0-8=2 → R2*=4
→ S3=S2+Q2-R2
S3=2+4-6=0 → This gives R3* = 8 and ;
→ S4=S3+Q3-R3
S4=8+0-8=0 → This gives R4* = 8
→ THEREFORE THE OPTIMAL RELEASE SEQUENCE IS (6,6,8,8) OR (8,4,8,8 ) .
AND BOTH GIVES OPTIMIZED OBJECTIVE FUNCTION OF f14 (S1) = 144 units.

7
2. A reservoir is planned both for gravity and lift irrigation through withdrawals from its
storage. The total storage available for both uses is limited to 5 units each year. It is
decided to limit the gravity irrigation withdrawal in a year to 4 units. If x1 is the
allocation of water to gravity irrigation and x2 is the allocation for lift irrigation, two
objectives are planned to be maximized and are expressed:
Z1 = 3x1 - 2x2
Z2 = -x1 + 4x2

a) Formulate multi objective planning model using weighting approach with weights for
gravity and lift irrigation withdrawals being w1 and w2 respectively. Plot the decision and
the objective space and determine the optimal share of withdrawals for gravity and lift
irrigations if

i. w1 =1 and w2 = 2,
ii. ii) w1 = 2 and w2 = 1

b) Formulate the problem using the Constraint Method

8
GIVEN
Reservior for uses of gravity and lift irrigation
Total reservior storage ,K=5 units
Gravity irrigation withdrawl is less or eqaul to 4
X1 = allocation of water to gravity irrigation
X2 = allocation of water to lift irrigation
REQUIRED
To formulate the model using weighting approach and find their results.if;
(i) w1 =1 & w2 = 2
(ii) w1 = 2 & w2 = 1
To formulate the model using constraint approach .
SOLUTION
Maximize Z = [Z1 (x) , Z2 (x)]
Z1 (x) = 3x1 -2x2
Z2 (x) = -x1 + 4x2
Subject to;
g1 (x): x1 + x2 ≤ 5
g2 (x): x1 ≤ 4
g3 (x): x2 ≥ 0
g4 (x): x1 ≥ 0

9
(i) Now when w1 = 1 & w2 = 2
Z = w1*z1+w2*Z 2
= Z 1 +2*Z 2
= (3x1 - 2x2 ) + 2*(-x1 + 4x2)
= x1 + 6x2
→ Now lets draw the decision space
x1 x2 z1 z2 Z= Z 1+2*Z 2 Z= 2*Z 1+1*Z 2
0 0 0 0 0 0
0 5 -10 20 30 0
4 1 10 0 10 20
4 0 12 -4 4 20
X2

(5,0) Z2
non-inferior set

(4,1)
Feasible Region
Z1
(0,0) (4,0)

→ And the objective space;


Z2

(-10,20)
non-inferior set

Feasible region (10,0) Z1


(0,0) (12,-4)

→ The new Z line has a slope of -1/6 in the decision space and testing at the corner points in
the decision space,Z will be maximum at (0,5) and its value is Zmax = 30.
and from the objective space,the line Z = Z 1 + 2*Z 2 has a slope of -1/2 and the value of Z
will be maximum at (Z1, Z2) = (-10, 20) in the objective space and Zmax = 30.

(ii) when w1 = 2 & w2 = 1


Z = w1*z1+w2*Z 2
= 2*Z 1+1*Z 2
= 2*(3x1 - 2x2 ) + 1*(-x1 + 4x2)
= 5x1

10
Hence,
In the decision space Z has a maximum value on the boundary at (4,1) and (4,0),equal to Zma
and in the objective space Z = 2*Z 1+ Z 2, Z is maximum at point (10,0) and (12,-4) ,
equal to Zmax = 20

(b) The problem above can be formulated to suit for constraint method as
Maximize :- Z 1 = 3x1 - 2x2
Subject to:- Z 2(x) = -x1 + 4x2 ≥ L2
g 1(x): x1+ x2 ≤ 5
g 2(x): x1 ≤ 4
g 3(x): x2 ≥ 0
g 4(x): x1 ≥ 0

11
→ if available water (reiease) is 2 units
ͮ→ User-1=X1 =0 units of water
ͮ→ User-2=X2 =0 units of water
ͮ→ User-3=X3 =2 units of water

→ Now we can set relationship between BENEFIT and RELEASE


RELEASEBENEFIT
0 0
2 14
4 24
6 32
8 40

b) RESORVIOR RELEASE POLICY (backwared recursion)

t=1 t=2 t=3 t=4


n=4 n=3 n=2 n=1
S1 = 0 S2 S3 S4
Q 1 =10 Q2= 2 Q3 = 8 Q 4 = 12

St +1 = St + Q t - Rt
ft n (St ) = Maximized net benefit
→ Stage-1
n= 1 S4
t=4 R4
Q 4 = 12
f4 1 (S4 ) = Max[B4 (R4 )]

12

You might also like