You are on page 1of 17

Optimisation and Decision-making coursework- ENGM072

Supervisor name: Franjo Ceclja


Name: Monawer Al-Hadid
Student number: 6400443

Faculty of Engineering
Renewable energy systems Engineering MSc
Table of Contents
List of figures: .................................................................................................................... 3
Introduction: ..................................................................................................................... 4
Aim: .................................................................................................................................. 4
Optimisation model: .......................................................................................................... 5
Profit: ................................................................................................................................ 6
Variables: .......................................................................................................................... 6
Parameters:....................................................................................................................... 6
Distances: .......................................................................................................................... 7
Scalars: .............................................................................................................................. 7
Cost: .................................................................................................................................. 8
Collection cost: .......................................................................................................................... 8
Cost of transportation: ............................................................................................................... 8
Storage cost: .............................................................................................................................. 8
Collection point constraint ....................................................................................................... 10
Storage capacity constraint ...................................................................................................... 10
Biorefinery constraint (1) ......................................................................................................... 10
Biorefinery constraint (2) ......................................................................................................... 10
Customer demand constraint ................................................................................................... 11
Results............................................................................................................................. 11
Scenario 1 by using technology 1 ............................................................................................. 11
Scenario 1 by using technology 1 and technology 2 .................................................................. 12
Scenario 2 by using technology 1 ............................................................................................. 14
Scenario 2 by using technology 1 and technology 2 .................................................................. 15
Conclusion: ...................................................................................................................... 16
References ....................................................................................................................... 17

List of tables:
Table 1:The following table represents the parameter, the availability of biomass at the
collection points, storage capacity at the three locations and the biorefineries capacity ....... 4
Table 2: Explanation of the set model ...................................................................................... 6
Table 3: The model parameters ............................................................................................... 7
Table 4: Distances between collection points (i) and storages (j) in miles. .............................. 7
Table 5: Distances between storages (j) and Biorefineries (k) in miles. ................................... 7
Table 6: The following table represents the scalars used in GAMS, their description and
values ....................................................................................................................................... 8
List of figures:
Figure 1: The total supply chain block diagram with distances between each location. ......... 5
Figure 2: Hessain matrix equation ............................................................................................ 9
Figure 3: finding the matrix definite by solving Eigenvalues (Surreylearn.surrey.ac.uk, 2019).
................................................................................................................................................ 10
Figure 4: Block diagram showing scenario 1 by using technology 1....................................... 11
figure 5: Block diagram showing scenario 1 by using technology 1 and technology 2………
Figure 6:Block diagram showing scenario 2 by using technology 1. ...................................... 14
Figure 7:Block diagram showing scenario 2 by using technology 1 and technology 2. .......... 15
Introduction:
A bio-fuel refining supply chain located in the United Kingdom consists of five collection
points of biomass, their locations are: Peterborough, King’s Lynn, Norwich, Ipswich and
Letchworth. The Wheat straw is used as the main source of raw material for producing
biofuel. Biofuel’s production mainly depends on Wheat straw as a source of raw material
Additionally, wheat straw costs £20 per tonne at the place of collection points. The wheat
straw is then transported from the collection points to the storage facilities located in
Thetford, Cambridge and Bury St. Edmunds, by using small trucks that charge £5 per mile
per tonne. Storing the wheat straw feedstock in all the storage locations is the same at a
rate of 5.25 per tonne. The biomass is then transported from the three storage facilities to
the biorefineries located in Cambridge and Colchester, by using large trucks at a cost of £2.5
per mile per tonne. The biorefinery in Cambridge uses technology 1 with a conversion rate
of biorefining of 30%. Meaning that every 1 tonne of wheat straw produces 0.3 tonne of
biofuel. On the other hand, the biorefinery in Colchester can use technology 1 or technology
2 with a conversion rate of 32%. The cost of biorefining process is £95 per tonne of biofuels
for both technologies and biofuel are sold at a price of £1025 per tonne.

Aim:
1. The aim of the first scenario is to analyse the percentage of utilisation of biomass in
the fields and storage capacity and maximizing the profit, by using the maximum
storage capacity in Cambridge of 50000 tonnes. Also, using the maximum capacity in
Colchester of 40000 tonnes if technology 1 is used or 35000 tonnes if technology
(1&2) is used.
2. The aim of the second scenario is to analyse the percentage of utilisation of biomass
in the fields, storage and bio-refinery capacities, if the customers demanded a total
of 28000 tonnes of biofuel.

Table 1:The following table represents the parameter, the availability of biomass at the
collection points, storage capacity at the three locations and the biorefineries capacity

COLLECTION POINT STORAGE CAPACITY BIOREFINERIES BIOFUEL


FEEDSTOCK (I) (J) (TONNES) CAPACITY (K) DEMAND (M)
AVAILABILITY (TONNES) (TONNES) (TONNES)
PETERBOROUGH Thetford Cambridge 28000
(25000) (50000) (50000)
KING’S LYNN
(35000)
NORWICH Cambridge Colchester
(25000) (20000) (40000) technology 1
IPSWICH
(30000)
LETCHWORTH Bury St. Edmunds Colchester
(35000) (40000) (35000) technology 2
Optimisation model:
It is defined as a type of mathematical model that try to minimise or maximise an objective
function by choosing the values from a range of allowed set and computing the value of the
function. The industrial supply chain’s main purpose is to allocate the most suitable path in
which the cost in minimised and the profit is maximised within the available capacity. A lot
of obstacles must be solved to make the function more profitable. Limitations of the
problem are stated by the constraint. The wheat straw’s availability and each of the
location’s maximum capacity are the main constraint of this case. Block diagrams are drawn
to describe the supply chain in a simplified way which helps detect the optimum routs to
minimize the cost of transportation.

Figure 1: The total supply chain block diagram with distances between each location.

The figure above, represents the total network of the bio refinery supply chain with all the
possible routes from the collection points to storages to biorefineries. The distances
between each location are stated above the route.
Profit:
The objective of the function is to maximise the profit. The profit can be calculated by
subtracting the revenue collected from selling the biofuel to the customers.
• The biorefinery is selling their biofuel at a price of 1025£ per tonne.
Revenue refers to the total sale price of the product multiplied be the quantity of product.
• Revenue= selling price * biofuel quantities sold to the customers
= 1025 * ∑ k,m (K,M).
Profit can be found by removing the total cost from the revenue.
• Profit= revenue – (Cost purchasing raw materials – Cost Bio-refining – Cost transportation – Cost storage)

Variables:
• (X I,j )= amount of wheat straw transferred from the collection point (i) to the
storage facilities (j) in tonnes.
• (Y j,k )= amount of wheat straw transferred from the storage facilities (j) to the
biorefineries (k) in tonnes.
• (Z k,m)= amount of Biofuel transferred from the biorefineries (k) to the customers (m)
in tonnes.
Table (2) shows the set used to outline the model. The location is stated with the set.

Table 2: Explanation of the set model

SET INTERPRETATION LOCATIONS


I Collection point Peterborough, Kingslynn,
Norwich, Ipswich,
Letchworth
J storages Thetford, Cambridge, Bury
St. Edmunds
K Biorefineries Cambridge, Colchester
M Customers

Parameters:
The parameters are identified as follows:
Þ The availability of biomass at different collection points, a(i).
Þ The maximum capacity of storing wheat straw at different locations, b(j).
Þ The maximum capacity of biorefineries, c(k).
Þ The customer’s need of biofuel, q(m).
The table below, represent the parameters description and the values of each parameter
measured in tonnes.

Table 3: The model parameters

PARAMETERS DESCRIPTION
A(I) availability of wheat straw at different
collection points
B(J) capacity of storing wheat straw at different
locations
C(K) Biorefinery capacities
Q(M) customer’s need of biofuel

Distances:
The wheat straw and biofuel were transported from a location to another by using small
and large trucks. As the cost is per mile per tonnes, it’s very important to state the distances
in mileage covered between each location. Google maps were used to gather the distances
between each location. An assumption was made by taking the fastest route, which consists
of less mileage to ensure minimizing the cost.

Table 4: Distances between collection points (i) and storages (j) in miles.

THETFORD CAMBRIDGE BURY ST. EDMUNDS


PETERBOROUGH 58.1 40.6 56.0
KING’S LYNN 31.4 45.6 40.8
NORWICH 30.8 64.5 43.6
IPSWICH 34.1 54.8 27.2
LETCHWORTH 58.3 27.4 52.9

Table 5: Distances between storages (j) and Biorefineries (k) in miles.

CAMBRIDGE COLCHESTER
THETFORD 34.5 49.6
CAMBRIDGE 0.0 48.8
BURY ST. EDMUNDS 29.0 30.9

Scalars:
Scalar statement is used to identify a GAMS parameter of dimensionality zero. There are no
connected sets, only one number associated with the parameter.
The constants used in the model equation are called scalars, they are helpful for calculating
the best route.
Table 6: The following table represents the scalars used in GAMS, their description and
values

SCALARSN DESCRIPTION VALUE


CONV1 Conversion rate of wheat straw to biofuel per tonne, by using technology 1 0.30
CONV2 Conversion rate of wheat straw to biofuel per tonne, by using technology 2 0.32
BMAS Cost of wheat straw at collection points per tonne 20.0
STOR Cost of storing wheat straw per tonne 5.25
REFI Cost of Biorefining one tonne of biomass 95
SELLINGY Price of biofuel sold to customers per tonne 1025
PRICE
VAN1 Transportation cost between collection points and storages using small trucks 5.00
per mile per tone
VAN 2 Transportation cost between collection points and storages using large trucks 2.50
per mile per tone
D Demand of biofuel for scenario 2 28000

Cost:
It’s very important to state all the costs involved in the supply chain phases, to improve the
problem. The costs involved in this supple chain process are as follows: collection cost,
storage cost, biorefining cost and transportation cost. By using GAMS software, the costs for
each section are calculated to solve the problem. The method behind calculating the cost
will be demonstrated in the sections below.

Collection cost:
It’s the cost of buying wheat straw from the collection point, the given value is £20 per
tonne. The total collection cost can be calculated by multiplying £20 by the total amount of
wheat straw transported from the collection point to the storages.
𝐶𝐶 = 𝐵𝑚𝑎𝑠∑𝑖, 𝑗(𝑥, 𝑗)

Cost of transportation:
The cost of transportation from the collection point to the storages is £5 per mile per tonne
by using small trucks, and £2.5 from the storages to the Biorefineries per mile per tonne by
using large trucks.
𝑇𝐶 = [𝑣𝑎𝑛1∑𝑖, 𝑗𝑥(𝑖, 𝑗) ∗ 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒_1] + [𝑣𝑎𝑛2∑𝑗, 𝑘𝑦(𝑗, 𝑘) ∗ 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒_2]

Storage cost:
The storage cost is defined as the cost of storing biomass. The cost is minimized in all three
storages to ensure profit maximization. Minimized cost represents a complex relationship
between energy requirements (x1), operational cost related to labour and equipment (x2)
and repayment of capital investment (x3).

Storagecost = x21 +x22 + x33 – x1 – x2 – x3 + 6;


Firstly, the cost of storage can be found by computing the gradient, which can be done by
solving the first derivative as shown below:
F’(x)= 2x1 + 2x2 + 2x3 – 1 – 1 – 1 = 0
2x1= 1;
2x2=1;
2x3=1;
X1=0.5;
X2=0.5;
X3=0.5;
By substituting into the original equation, we get the minimized cost:
Storagecost = 0.52 + 0.52 + 0.52 – 0.5 – 0.5 + 6.0 = £5.25

The same answer was obtained from GAMS software, the file is attached with the main
GAMS. But, in order to verify that the results we obtained is the minimum localiser, the
second derivative must be calculated by using Hessian matrix.

Figure 2: Hessain matrix equation

Solving hessian matrix to obtain the determinant:


> ? @(AB,AC,AD)E
>AB
𝑏 = 2𝑥1;
> ? @(AB,AC,AD)
>AC
𝑏 = 2𝑥2;
> ? @(AB,AC,AD)
>AD
𝑐 = 2𝑥3;

After finding the determinant, Eigen values must be calculated to ensure that the matrix is
positive. The storage cost can be confirmed as the local minimizer, if the matrix is positive.

B= 3*3 matrix, where (I) is the matrix identity and 𝜆 is the eigen value.
Figure 3: finding the matrix definite by solving Eigenvalues (Surreylearn.surrey.ac.uk, 2019).

So 𝜆 = 2, by proving that all Eigen values are positive, the storage cost is now confirmed to
be the minimum cost.

Constraints
constraints must be included when an optimisation model is constructed. Constraints are
mainly used to give a region of feasibility, in order to achieve the maximum value, the
model must run within these limits. In this case, the constraints are the availability of wheat
straw in the collection points, the maximum capacity of each storage and balancing the
distribution of wheat straw.

Collection point constraint


The first limit that must be taken into consideration is the collection constraint, which
controls the availability of wheat straw in each collection point. This constraint shows that
the sum of wheat straw transferred from the collection points to the storages must be less
than or equal to the availability of wheat straw in the collection point.

𝑐𝑜𝑙𝑙𝑒𝑐𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡 = ∑𝑗 𝑥(𝑖, 𝑗) ≤ 𝑎(𝑖);

Storage capacity constraint


This constraint states that the amount of wheat straw transferred from the collection
location to the storage should not more the maximum storage capacity.
𝑠𝑡𝑜𝑟𝑎𝑔𝑒 𝑐𝑎𝑝𝑐𝑎𝑖𝑡𝑦 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡 = ∑𝑖𝑥(𝑖, 𝑗) ≤ 𝑏(𝑗);

Biorefinery constraint (1)


This constraint is for scenario 1, where the amount of wheat straw transferred from the
storages to the biorefinery must be less than the biorefinery maximum capacity.
𝑠𝑒𝑐𝑛𝑎𝑟𝑖𝑜1 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡 = ∑𝑗 𝑦(𝑗, 𝑘) ≤ 𝑐(𝑘);

Biorefinery constraint (2)


This constraint is for scenario 2, where the amount of wheat straw transferred from the
storages to the biorefinery must be equal to the maximum capacity of biorefinery.
𝑠𝑐𝑒𝑛𝑎𝑟𝑖𝑜2 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡 = ∑𝑗 𝑦(𝑗, 𝑘) = 𝑐𝑐(𝑘);
Customer demand constraint
This constraint is for scenario 2, where the amount of biofuel sold to the customers must be
equal to the customers demand which is 28000 tonnes.
𝑑𝑒𝑚𝑎𝑛𝑑 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡 = ∑𝑘 𝑧(𝑘, 𝑚) = 𝑞(𝑚);

Results
Scenario 1 by using technology 1
All the following scenarios were solved by using CONOPT, which can solve linear and
nonlinear programming. It’s also known for its ability to self-tune.
The maximized profit by using GAMS software is £3,940,000. The equation to calculate the
profit is given below:

Profit1= e1= [selling price*∑j,k y(j,k)*conversion1(k)] – [refi*∑j,k y(j,k)*coversion_1(k)] –


[van1*∑I,j x(I,j)*distance_1(I,j)] – [van2*∑j,k y(j,k)*distance_2(j,k)] – [stor*∑I,j x(I,j)] –
[bmas*∑I,j x(I,j)].

Figure 4 shows a graphical representation of the results by using GAMS software for
scenario 1 by using technology 1.

Figure 4: Block diagram showing scenario 1 by using technology 1


According to figure (3), Peterborough did not contribute in distributing the wheat straw
supply to any of the storages due to high mileage that It needs to cover. On the other hand,
Ipswich contributed all its supply to Bury St. Edmunds due to the short distance between
the collection point and the storage. Ipswich is only 27.2 miles away from Bury St. Edmunds.
Letchworth contributed 57% of its supply to Cambridge due to the close distance between
them. Moreover, Norwich contributed all of its wheat straw supply to Thetford which is
considered the nearest when compared to Cambridge and Bury St. Edmunds. Furthermore,
King’s Lynn contributed 14.3% of its supply to Thetford which is 31.4 miles away from it.
Thetford received 30000 tonnes of wheat straw which represents 60% of its capacity.
Cambridge received its full capacity of wheat straw which is 20000 tonnes and Bury St.
Edmunds received 30000 tonnes of wheat straw which contribute to 75% of its maximum
capacity. Moving to the second part of the supply chain, which is from the storages to the
biorefinery. All the amount of wheat straw transferred to the storages was transferred to
the biorefineries. Cambridge biorefinery received 50000 tonnes of wheat straw which is the
maximum capacity. Whereas, Colchester received only 30000 tonnes which contribute to
70% of its capacity when technology 1 is used.

To verify the results of GAMS software, the values must be substituted in the profit
equation and the same answer must be obtained.

Profit1=e1= [182.25*5000] + [179.25*25000] + [161.25*30000] + [162.25*20000] –


[30000*192.75] – [20000*279] – [30000*201.75]= £3,940,000.
By getting the same value as the GAMS software we can guarantee that the software is
accurate.

Scenario 1 by using technology 1 and technology 2


The maximized profit by using GAMS software is £4,454,750. The following equation shows
how the profit was calculated:

Profit2= e2=[selling price*∑j,k y(j,k)*conversion2(k)] – [refi*∑j,k y(j,k)*conversion2(k)] –


[van1*∑I,j x(I,j)*distance_1(I,j)] – [van2*∑j,k y(j,k)*distance_2(j,k)] – [stor*∑I,j x(I,j)] –
[bmas*∑I,j x(I,j)].

the figure below shows a graphical representation of the results by using GAMS software for
scenario 1 by using technology 1 and technology 2.
Figure 5: Block diagram showing scenario 1 by using technology 1 and technology 2.

By referring to figure (4), Peterborough did not contribute in the distribution of wheat straw
for the same reason mentioned in scenario 1 by using technology 1. King’s Lynn distribution
of wheat straw increased by 50% from 5000 tonnes to 10000 tonnes, when using
technology (1 and 2) when compared to using technology 1 only. King’s Lyon contributed
about 28.6% of its maximum capacity to transfer wheat straw from the collection points to
the storages. Norwich, Ipswich and Letchworth contributed to the same amounts as
scenario 1 using technology 1. The capacity of Thetford increased by 5000 tonnes when
compared to scenario 1 using technology 1, which represents 70% of its maximum capacity.
Whereas, Cambridge reached its maximum capacity of 20000 tonnes. Moreover, Bury St.
Edmunds received the same amount of wheat straw as the above scenario. The second part
of this supply chain, which represents the transportation from the storages to the
biorefineries. Cambridge biorefinery received 50000 tonnes which translates to the full
capacity that it can take. On the other hand, Colchester received 5000 tonnes more which
contributes to reaching the maximum capacity of raw materials.

Profit value obtained by GAMS can be verified when the wheat straw capacity values are
substituted in the profit equation.

Profit2=e2= [182.25*10000] + [179.25*25000] + [161.25*30000] + [162.25*20000] –


[30000*192.75] – [5000*173.6] – [20000*279] – [30000*220.35] = £4,454,750.
Scenario 2 by using technology 1
The optimized profit by using GAMS software and the optimum route is £3,667,500. The
following equation shows how the profit was calculated:

Profit3= e3= [selling price*∑j,k y(j,k)*conversion1(k)] – [refi*∑j,k y(j,k)*coversion1(k)] –


[van1*∑I,j x(I,j)*distance_1(I,j)] – [van2*∑j,k y(j,k)*distance_2(j,k)] – [stor*∑I,j x(I,j)] –
[bmas*∑I,j x(I,j)].

Figure 6:Block diagram showing scenario 2 by using technology 1.

By referring to the above figure, it shows that King’s Lynn transferred about 43% of its
maximum supply to Thetford due to the short distance between them when compared to
Cambridge and Bury St. Edmunds. The distance between Norwich and Cambridge is 64.5
miles, whereas, the distance between Norwich and Bury St. Edmunds is 43.6 miles. To
minimize the cost and maximize the profit, all Norwich supply was transferred to Thetford
due to the short distance of 30.8 miles between them. Ipswich transferred all of its supply to
Bury St. Edmunds. Moreover, Letchworth transferred 57.1% of its supply to Cambridge,
considering the closer distance between them. In comparison to both Thetford and Bury St.
Edmunds, where in accordance to figure 1, it appears to be further by around 23 miles.
Cambridge biorefinery reached its maximum capacity of 20000 tonnes. Furthermore,
Thetford and Bury St. Edmunds reached 80% and 75% of their maximum capacity
respectively.

GAMS answer can be verified by solving the mathematical equation, given below:
Profit3 =e3= [10000*182.25] + [25000*179.25] + [20000*162.25] + [30000*161.25] –
[30000*192.75] – [10000*155] – [20000*279] – [30000*201.75] = £3,667,500.
The same profit answer is obtained by solving the mathematical equation, GAMS profit
value is verified and trusted.
Scenario 2 by using technology 1 and technology 2
By using GAMS software, the optimized profit is £4,454,750. The following equation shows
how the profit was calculated.

Profit4=e4=[selling price*∑j,k y(j,k)*conversion2(k)] – [refi*∑j,k y(j,k)*coversion2(k)] –


[van1*∑I,j x(I,j)*distance_1(I,j)] – [van2*∑j,k y(j,k)*distance_2(j,k)] – [stor*∑I,j x(I,j)] –
[bmas*∑I,j x(I,j)].

Figure 7:Block diagram showing scenario 2 by using technology 1 and technology 2.

According to figure (6), king’s Lynn transferred 10000 tonnes to Thetford which represents
28.6% of its supply. By using a mixture of technology 1 and 2 in this scenario, the supply
from King’s Lynn to Thetford decreased by 5000 tonnes, when compared to using
technology 1 only. Norwich, Ipswich and Letchworth transferred the same amount of supply
of about 100%, 85.7% and 57.1% of their maximum capacity. All wheat straw transferred to
the storages must be transferred to the biorefineries. Furthermore, Thetford, Cambridge
and Bury St. Edmunds stored 70%, 100% and 75% of their maximum capacity, respectively.
Moving to the last part of the supply chain, Cambridge and Colchester received 100% of
their maximum capacity.

Profit4=e4= [10000*182.25] + [25000*179.25] + [20000*162.25] + [30000*161.25] –


[30000*192.75] – [5000*173.6] – [20000*279] – [30000*220.35] = £4,454,750.
Conclusion:
To conclude, the objective of scenario 1 has been fulfilled, as the utilization percentages
where analysed for technology 1 and technology (1&2). For scenario 1, the optimum route
has been found and the profit has been maximized to reach a value of £4,454,750 by using a
combination of technology 1 and 2. On the other hand, the maximized profit obtained by
using technology 1 only is £3,940,000. Moreover, the maximum profits for scenario 2
obtained by using technology 1 and technology (1&2) are £3,667,500 and £4,454,750,
respectively. Whereas, both technology 1 and technology (1&2) in scenario 2 are infeasible,
as the total customer demand did not reach the maximum demand of 28000. Furthermore,
by using different sensitivity check the maximized profit can be improved, as the cost
causing most loss in the profit can be minimized.
References:

• Surreylearn.surrey.ac.uk. (2019). [online] Available at:


https://surreylearn.surrey.ac.uk/d2l/le/content/189595/viewContent/1639984/View [Accessed 7
Dec. 2019].

You might also like