IOSR Journal of Mathematics (IOSR-JM)

e-ISSN: 2278-5728, p-ISSN: 2319-765X. Volume 19, Issue 1 Ser. II (Jan. – Feb. 2023), PP 29-34
www.iosrjournals.org

Approximation Method for Solving TwoDimensional

Cutting-Stock Problems
Abdul Rahman1, Ihda Hasbiyati2, M. D. H. Gamal3
1
(Department of Mathematics, Faculty of Mathematics and Science, Universityof Riau,Indonesia)
2
(Department of Mathematics, Faculty of Mathematics and Science, University of Riau,Indonesia)
3
(Department of Mathematics, Faculty of Mathematics and Science, University of Riau,Indonesia)

Abstract:
The problem of cutting two-dimensional stock is a problem where the cutting pattern considers the length and
width of a rectangular stock. The objective is to minimize the wastage which is equivalent to minimizing the
number of stocks used. This article discusses the solution to the problem of cutting two-dimensional stocks
assuming that the supply is unlimited, and demands requested when cutting cannot be rotated. Then, the column
generating technique is implemented to obtain a knapsack problem, and the knapsack is solved by dynamic
programming method. The result obtained is a best combination of cutting patterns, so it minimizes the cutting
wastages.
Key Word: Column generating; Cutting pattern; Knapsack problem; Rectangular stock; Two dimensional
cutting-stock.
----------------------------------------------------------------------------------------------------------------------------- ---------
Date of Submission: 10-02-2023 Date of Acceptance: 22-02-2023
----------------------------------------------------------------------------------------------------------------------------- ----------

I. Introduction
Linear programming is a technique from operations research to find solutions to optimization problems
using linear equations and inequalities in order to obtain optimal solutions taking into account the existing
constraints. In solving a problem, linear programming uses a mathematical model. This model presents the form
and arrangement of the problem to be solved. The linear programming develops and provides a strong impetus
for industrial progress. One of its applications is to solve the cutting-stock problems. The cutting-stock problem
is one of the optimization problems that basically reduces a linear programming into a linear programming with
integer values.
The problem is called as a cutting-stock if there are limited demands or orders with various sizes while
ignoring the other constraints. The demands are fulfilled by cutting the stocks with one or more standard sizes.
To form a linear programming as the main problem, the possible objective function is to minimize the total
amount of stocks with standard size cut.
One dimensional cutting-stock problemconsiders only one type of cutting, namely length or width.
Regarding one dimensional cutting-stock, Gilmore and Gomory[1]use a linear programming approach where
stocks are available in various standard sizes 𝐿1 , 𝐿2 , … , 𝐿𝑘 using the cost function and then solve it using the
dynamic programming method. Similarly, Winston[2]discusses a linear programming approach to solve the
onedimensional cutting-stock where the cutting patterns are generated using the column generating technique
with the help of the branch-and-bound method. In addition, Wongprakornkul and Charnsethikul[3]solve the
problem of cutting one-dimensional stock with limited demands and objective capacity planning.
The two dimensional cutting-stock problem is an issue where the cutting patterns used consider the
length and width of a rectangular stock. The previous papers on these two dimension cases include[4], [5], [6],
[7], [8] and [9]. The problem of cutting two-dimensional stock also uses a column generation technique which is
solved using a knapsack problem. Several articles have been written regarding the use of knapsack in this two-
dimensional problem[10], [11], [12], [13] and [14].
This article discusses how to solve the problem of cutting two-dimensional stock with the assumption
that cutting is only done on one type of stock measuring length, while width and height are ignored. Then the
availability of supply is unlimited, and demands are fulfilled by cutting the standard stocks where the cutting
cannot be rotated. Then the column generation technique is used to obtain the knapsack shape, and the knapsack
problem is solved by considering the knapsack problem in one dimension with the dynamic programming
method. The first solves the knapsack problem by making the problem case-by-case basis by using a width
constraint. Then, the value of the objective function of each case is used to solve the knapsack problem with
length constraints by making the knapsack problem case-by-case.

DOI: 10.9790/5728-1901022934 www.iosrjournals.org 29 | Page

 Approximation Method for Solving Two Dimensional Cutting- Stock Problems

II. Methods
Column Generation Technique
Column generation is an iterative algorithm that starts with a small set of initial patterns, and then
cleverly chooses new columns to add to the main cutting-stock problem so that we find the optimal solution
without having to enumerate every column. Columns in the cutting-stock mathematical model indicate the
cutting patterns. The column generation method takes advantage of the effectiveness of the revised simplex
method in the cutting-stock problems, so that many of the processing steps refer to the revised simplex method.
Let us consider the problem of cutting iron plate with standard length 𝐿 and standard width 𝑊 which
will be cut into length 𝑙𝑖 and width 𝑤𝑖 for 𝑖 = 1, 2, … , 𝑚 with the number of orders as many as 𝑏𝑖 . Variable 𝑥𝑗
denotes the number of standard stocks cut according to pattern or activity 𝑗, 𝑎𝑖𝑗 is the number of standard stocks
with length 𝑙𝑖 and width𝑤𝑖 resulting from activity 𝑗. So the general form of the cutting-stock problem is
min 𝑧 = 𝑥1 + 𝑥2 + ⋯ + 𝑥𝑛 ,
subject to 𝑎𝑖1 𝑥1 + 𝑎𝑖2 𝑥2 + ⋯ + 𝑎𝑖𝑛 𝑥𝑛 ≥ 𝑏𝑖 , 𝑖 = 1, 2, … , 𝑚, (1)
𝑥𝑗 ≥ 0 and integer, 𝑗 = 1, 2, … , 𝑛.
Equation (1) can be written in matrix form and in standard form as follows:
min 𝑧 = 𝑐𝐵𝑇 𝑥𝐵 + 𝑐𝐵𝑇 𝑥𝑁 ,
subject to𝐵𝑥𝑁 + 𝑁𝑥𝑁 = 𝑏,
𝑥𝐵 , 𝑥𝑁 ≥ 0 and integer
where 𝑥𝐵 and 𝑥𝑁 are basic variables and nonbasic variables, 𝐵 dan 𝑁 are the column matrices corresponding to
the variables 𝑥𝐵 and 𝑥𝑁 . During the simplex method iterations, suppose that the associated basis is 𝐵 =
𝐴1 , 𝐴2 , … , 𝐴𝑚 where 𝐴𝑖 is a column vector of dimension 𝑚, 𝑐𝐵𝑇 = (𝑐1 , 𝑐2 , … , 𝑐𝑚 ) is the coefficient of the
objective function that corresponds to 𝐴1 , 𝐴2 , … , 𝐴𝑚 . Then from the linear programming, the promising cutting
pattern 𝑗 is the one whose reduced cost
𝑧𝑗 − 𝑐𝑗 = 𝑐𝐵𝑇 𝐵 −1 𝐴𝑗 − 𝑐𝑗
is positive for the minimization problem, where
𝑇
𝐴𝑗 = 𝑎1𝑗 , 𝑎2𝑗 , … , 𝑎𝑚𝑗
is a vector indicating the number of cuts with 𝑙𝑖 and 𝑤𝑖 resulting from the cutting pattern 𝑗.
The pattern that gives an advantage is the one whose value of 𝑧𝑗 − 𝑐𝑗 is positive for the minimization
problem. For this reason, the column generation method is needed, namely a technique to obtain a most
promising column or pattern. In this case, it is equivalent to solving the subproblem
max 𝑣 = 𝑑1 𝑦1 + 𝑑2 𝑦2 + ⋯ + 𝑑𝑚 𝑦𝑚 − 1
subject to 𝑟1 𝑦1 + 𝑟2 𝑦2 + ⋯ + 𝑟𝑚 𝑦𝑚 ≤ 𝑅
(
2)
𝑦𝑖 ≥ 0 and integer,
where 𝑑𝑖 is the dual value of 𝑖 from 𝑐𝐵𝑇 𝐵−1 , 𝑦𝑖 is nonnegative integer, 𝑟𝑖 is amount for order 𝑖 and 𝑅is standard
size. Problem (2) is also called as knapsack problem.

Knapsack Problem
The knapsack problem is an integer linear programming problem with one constraint. The knapsack
problem is used to find the best solution for all possible items to be packed into the sack. For this reason, the
dynamic programming method is used in solving the knapsack subproblem. Taha[15]explaines that the dynamic
programming is to determine the optimal solution of a multivariable problem by decomposing it into several
stages, each stage consisting of a single variable subproblem. The following are the steps in solving the
knapsack problem using the dynamic programming method in the two-dimensional cutting stock problem:
1. Sorting the length size of the demand from the smallest to the largest and the width of the demand follows
the order of the length size.
2. Solving the knapsack problem using the width constraint. The following is the linear programming form of
the knapsack problem with initial truncation using the width constraint:
max 𝑣 = 𝑑1 𝑦1 + 𝑑2 𝑦2 + ⋯ + 𝑑𝑚 𝑦𝑚 − 1,
subject to𝑤1 𝑦1 + 𝑤2 𝑦2 + ⋯ + 𝑤𝑚 𝑦𝑚 ≤ 𝑊, (3)
𝑦𝑖 ≥ 0 and integer.
3. Solving subproblem (3) by dividing case by case, namely 𝑤1 𝑦1 ≤ 𝑊, 𝑤1 𝑦1 + 𝑤2 𝑦2 ≤ 𝑊, … , 𝑤1 𝑦1 +
𝑤2 𝑦2 + ⋯ + 𝑤𝑚 𝑦𝑚 ≤ 𝑊with dynamic programming method.
4. Obtaining the objective function value from each case, for example with 𝑒1 , 𝑒2 , … , 𝑒𝑛 . The objective
function values 𝑒1 , 𝑒2 , … , 𝑒𝑛 are used to solve the knapsack problem by using the length constraint. The
following is the linear programming form of the knapsack problem using the length constraint:

DOI: 10.9790/5728-1901022934 www.iosrjournals.org 30 | Page

 Approximation Method for Solving Two Dimensional Cutting- Stock Problems

max 𝑣 = 𝑒1 𝑦1 + 𝑒𝑦2 + ⋯ + 𝑒𝑚 𝑦𝑚 − 1,
subject to𝑙1 𝑦1 + 𝑙2 𝑦2 + ⋯ + 𝑙𝑚 𝑦𝑚 ≤ 𝐿, (4)
𝑦𝑖 ≥ 0 and integer.
5. Solving subproblem (4) by dividing case by case, namely 𝑙1 ≤ 𝐿, 𝑙1 𝑦1 + 𝑙2 𝑦2 ≤ 𝐿, … , 𝑙1 𝑦1 + 𝑙2 𝑦2 + ⋯ +
𝑙𝑚 𝑦𝑚 ≤ 𝐿 with dynamic programming method.

III. Result
From the problem of cutting-stock, for example, a iron plate company produces iron plate in feet (ft) with
the length of 12ft and the width of 10ft. Then, there are some orders requested as shown in Table no 1.

Tabel no 1: Order sample

Order Length used Width used Number of order
(𝑖) (ft) (ft) (sheets)

1 3 4 22
2 4 5 14
3 7 3 18

The initial basis variables (initial basis) is obtained by selecting from pure truncation patterns. So from the pure
cutting pattern, the initial basis for lengths 3ft × 4ft, 4ft × 5ft, and 7ft × 3ft yields
1
0 0
8 0 0 8
1
𝐵0 = 0 6 0 and 𝐵0−1 = 0 6 0 ,
0 0 3 1
0 0
3
The dual value is obtained from the following calculation:
1
0 0
8
1 1 1 1
𝑐𝐵 𝐵0−1 = 1 1 1 0 0 = ,
6 8 6 3
1
0 0
3
for the new basis, a pattern verified by 𝑦1 , 𝑦2 dan 𝑦3 . The value of 𝑧𝑗 − 𝑐𝑗 is obtained as follows:
𝑦1
−1 𝑦
1 1 1
𝑐𝐵 𝐵0 2 − 1 = 𝑦1 + 𝑦2 + 𝑦3 − 1,
𝑦3 8 6 3
where 𝑦1 , 𝑦2 and 𝑦3 cannot exceed 10m, 𝑦1 , 𝑦 and 𝑦3 are nonnegative integers and patterns 𝑦1 , 𝑦2 and 𝑦3 must
satisfy
4𝑦1 + 5 𝑦2 + 3𝑦3 ≥ 10,
𝑦𝑖 ≥ 0 and integer.
To get a profitable pattern is by solving the equivalent knapsack problem
1 1 1
max 𝑣 = 𝑦1 + 𝑦2 + 𝑦3 − 1,
8 6 3
subject to 4𝑦1 + 5𝑦2 + 3𝑦3 ≤ 10, (5)
𝑦𝑖 ≥ 0 and integer.

The knapsack problem from subproblem (5) is a knapsack problem using the width constraint. This problem is
solved by using dynamic programming method.
To solve this problem, starting from the initial stage(𝑖 = 1) to the final stage or moving forward.
Case 1: 4𝑦1 ≤ 10.
By using the dynamic programming method, the optimal solution for case 4𝑦1 ≤ 10 is 𝑦1 = 2 and
2
𝑒1 = 𝑣 = .
8
Case 2: 4𝑦1 + 5𝑦2 ≤ 10.
By using the dynamic programming method, the optimal solution for case 4𝑦1 + 5𝑦2 ≤ 10 is
2
𝑦1 = 0, 𝑦2 = 2 and 𝑒2 = 𝑣 = .
6
Case 3: 4𝑦1 + 5𝑦2 + 3𝑦3 ≤ 10.
By using the dynamic programming method, the optimal solution for case 4𝑦1 + 5𝑦2 + 3𝑦3 ≤ 10 is
3
𝑦1 = 0, 𝑦2 = 0, 𝑦3 = 3 and 𝑒3 = 𝑣 = = 1.
3

DOI: 10.9790/5728-1901022934 www.iosrjournals.org 31 | Page

 Approximation Method for Solving Two Dimensional Cutting- Stock Problems

Then, solving the knapsack problem using the length constraint and the objective function value of each
case obtained from solving the knapsack problem using the width constraint, so that the equation of the
knapsack problem is as follows
2 1
max 𝑣 = 𝑦1 + 𝑦2 + 1𝑦3 − 1,
8 3
subject to 3𝑦1 + 4𝑦2 + 7𝑦3 ≤ 12, (6)
𝑦𝑖 ≥ 0 and integer.

To solve equation (6), starting from the initial stage (𝑖 = 1) to the final stage or moving forward.
Case 1: 3𝑦1 ≤ 12.
By using the dynamic programming method, the optimal solution for case 3𝑦1 ≤ 12 is 𝑦1 = 4 dan
𝑣 = 𝑣1 = 0.
Case 2: 3𝑦1 + 4𝑦2 ≤ 12.
By using the dynamic programming method, the optimal solution for case 3𝑦1 + 4𝑦2 ≤ 12 is 𝑦1 =
0, 𝑦2 = 3 and 𝑣 = 𝑣2 = 0.
Case 3: 3𝑦1 + 4𝑦2 + 7𝑦3 ≤ 12.
By using the dynamic programming method, the optimal solution for case 3𝑦1 + 4𝑦2 + 7𝑦3 ≤ 12 is
1
𝑦1 = 0, 𝑦2 = 1, 𝑦3 = 1 and 𝑣 = 𝑣3 = .
3
from cases 1, 2 and 3 which are obtained by using the length constraint, the maximum value of the objective
1
function v is found in the 3rd case with 𝑣 = 𝑣3 = , so that the optimal solution of the knapsack problem from
3
1
equation (7) is obtained by an integer solution, namely 𝑦1 = 0, 𝑦2 = 1 and 𝑦3 = 1which results in 𝑣 = . Then,
3
because the knapsack solution with length constraints corresponds to the knapsack solution with width
constraints 𝑦1 = 0, 𝑦2 = 2 and 𝑦3 = 3, so that a new value is obtained, 𝑦1 = 0 × 0 = 0 , 𝑦2 = 1 × 2 = 2 and
𝑦3 = 1 × 3 = 3.
1
The cutting pattern 𝑦1 = 0 , 𝑦2 = 2 and 𝑦3 = 3is not optimal yet, because the value of 𝑣 = . Next, a
3
pattern will be sought that will provide an advantage by including the cuts that have been obtained into the
basis, for this it is necessary to form a new right-hand side and a new column
1
0 0
0 8 0 0
−1 1 1
New column of 𝑥4 = 𝐵0 2 = 0 6 0 2 = 3 ,
3 1 3 1
0 0
3
1
0 0 11
8 18 4
1
New right side= 𝐵0−1 𝑏 = 0 6
0 14 = 7 .
1 22 3
0 0 6
3
The ratio test shows that 𝑥4 enters a new base in row 3, the new basis variable is BV = 𝑥1 , 𝑥2 , 𝑥4 ,
using the inverse product, we get
1 1
0 0 0 0
1 0 0 8 8
−1 −1 1 1 1 1
𝐵1 = 𝐸0 𝐵0 = 0 1 − 3 0 6 0 = 0 6 − 9 .
0 0 1 0 0 1 0 0
1
3 3
Now
1
0 0
8
1 1 1 1 2
𝑐𝐵 𝐵1−1 = 1 1 1 0 6
−
9
= 8 6 9
.
1
0 0
3
Again, the column generation technique is used to determine the pattern that will enter the base. For the
current dual value 𝑐𝐵 𝐵1−1 , a pattern defined by 𝑦1 , 𝑦2 and 𝑦3 determines the 𝑧𝑗 − 𝑐𝑗 value to
𝑦1
−1 𝑦 1 1 2
𝑐𝐵 𝐵1 2 − 1 = 𝑦1 + 𝑦2 + 𝑦3 − 1.
8 6 9
𝑦3
The equivalent knapsack problem is
1 1 2
max 𝑣 = 𝑦1 + 𝑦2 + 𝑦3 − 1,
8 6 9
subject to 4𝑦1 + 5 𝑦2 + 3𝑦3 ≤ 10, (7)

DOI: 10.9790/5728-1901022934 www.iosrjournals.org 32 | Page

 Approximation Method for Solving Two Dimensional Cutting- Stock Problems

𝑦𝑖 ≥ 0 and integer.
The knapsack problem from equation (7) is solved by by using dynamic programming method using
width as a constraint.
To solve this problem, starting from the initial stage (𝑖 = 1) to the final stage or moving forward.
Case 1: 4𝑦1 ≤ 10.
By using the dynamic programming method, the optimal solution for case 4𝑦1 ≤ 10 is 𝑦1 = 2 and
2
𝑒1 = 𝑣 = .
8
Case 2: 4𝑦1 + 5 𝑦2 ≤ 10.
By using the dynamic programming method, the optimal solution for case 4𝑦1 + 5 𝑦2 ≤ 10 is
2 1
𝑦1 = 0, 22 = 2 and 𝑒2 = 𝑣 = = .
6 3
Case 3:4𝑦1 + 5 𝑦2 + 3𝑦3 ≤ 10.
By using the dynamic programming method, the optimal solution for case 4𝑦1 + 5 𝑦2 + 3𝑦3 ≤ 10 is
6
𝑦1 = 0, 𝑦2 = 0, 𝑦3 = 3 and 𝑒3 = 𝑣 = .
9
Then, solving the knapsack problem using the length constraint and the objective function value of each
case obtained from solving the knapsack problem using the width constraint, so that the equation of the
knapsack problem is as follows
2 1 6
max 𝑣 = 𝑦1 + 𝑦2 + 𝑦3 − 1,
8 3 9
subject to 3𝑦1 + 4𝑦2 + 7𝑦3 ≤ 12, (8)
𝑦𝑖 ≥ 0 and integer.
To solve equation (8), starting from the initial stage (𝑖 = 1) to the final stage or moving forward.
Case 1: 3𝑦1 ≤ 12.
By using the dynamic programming method, the optimal solution for case 3𝑦1 ≤ 12 is 𝑦1 = 4 and
𝑣 = 𝑣1 = 0.
Case 2: 3𝑦1 + 4𝑦2 ≤ 12.
By using the dynamic programming method, the optimal solution for case 3𝑦1 + 4𝑦2 ≤ 12 is 𝑦1 =
0, 𝑦2 = 3 and 𝑣 = 𝑣2 = 0.
Case 3: 3𝑦1 + 4𝑦2 + 7𝑦3 ≤ 12.
By using the dynamic programming method, the optimal solution for case 3𝑦1 + 4𝑦2 + 7𝑦3 ≤ 12 is
𝑦1 = 0, 𝑦2 = 1, 𝑦3 = 1 and 𝑣 = 𝑣3 = 0.
from cases 1, 2 and 3 which are obtained by using the length constraint, the maximum value of the objective
function v is found in the 3rd case with 𝑣 = 0, so that the optimal solution of the knapsack problem of equation
(8) is obtained by an integer solution, namely 𝑦1 = 0, 𝑦2 = 1and 𝑦3 = 1 which results 𝑣 = 0. Then, because
the knapsack solution with length constraints corresponds to the knapsack solution with width constraints
𝑦1 = 0, 𝑦2 = 2 dan 𝑦3 = 3, so the new values 𝑦1 , 𝑦2 and 𝑦3 obtained are 𝑦1 = 0 × 0 = 0, 𝑦2 = 1 × 2 = 2 and
𝑦3 = 1 × 3 = 3.
Because the solution 𝑣 = 0, this means that there are no more profitable patterns if it is included in the
basis because the basic variable is already optimal. The basic variable obtained is {𝑥1 , 𝑥2 , 𝑥4 }. To determine the
basic variables in the optimal solution, look for the value of the right-hand side as follows:

1
0 0 11
8
1 1 22 4
𝐵1−1 𝑏 = 0 − 14 = 1 .
6 9 18
1 3
0 0 6
3

IV. Discussion
The optimal solution for the cuttting-stock problem above is 𝑥1 = 11/4, 𝑥2 = 1/3and 𝑥4 = 6. Integer
solutions are obtained by rounding up, namely 𝑥1 = 3, 𝑥2 = 1and 𝑥4 = 6. A Request of 22 sheets for size
3ft×4ft, 14 sheets for size 4ft ×3 ft and 18 sheets for size 7ft ×3ft can be fulfilled by cutting iron plate with the
standard size 12ft ×10ft. Cutting process is done by adjusting the cutting knife. Initial cutting of the iron plate
from the long side then to the wide side. In this case, the items demande of deduction cannot be rotated so that
they are obtained as follows:
1. Cut 3 sheets of standard size iron plate 12 ft x 10 ft. First cut 12 ft with a length of 3 ft as many as 4 sheets
then 10 ft with a width of 4 ft as many as 2 sheets so obtained 8 sheets of size 3 ft x 4 ft for each standard
size of 12 ft x 10 ft. The result in total is 24 sheets size 3 ft x 4 ft.

DOI: 10.9790/5728-1901022934 www.iosrjournals.org 33 | Page

 Approximation Method for Solving Two Dimensional Cutting- Stock Problems

2. Cut 1 sheet of standard size iron plate 12 ft x 10 ft. First cut 12 ft with a length of 4 ft as many as 3 sheets
then 10 ft with a width of 5 ft as many as 2 sheets so obtained 6 sheets of size 4 ft x 5 ft for each standard
size of 12 ft x 10 ft. The result in total is 6 sheets size 4 ft x 5 ft.
3. Cut 3 sheets of standard size iron plate 12 ft x 10 ft. First cut 12 ft with a length of 7 ft as many as 1 piece
and 4 ft as many as 1 sheet then 10 ft with a width of 3 ft as many as 3 sheets and 5 ft as many as 2 sheets
so obtained 3 sheets of size 7 ft x 3 ft and 2 sheets of size 4 ft x 5 ft for each standard size of 12 ft x 10 ft.
The result in total is 18 sheets size 7 ft x 3 ft and 12 sheets size 4 ft x 5ft.
Based on the solution above, deviation of the required number of iron plate is about 25% from the
result obtained from complete enumeration.

V. Conclusion
The problem of cutting a two-dimensional stock is a problem that considers length and width when
cutting. Many patterns can be formed when making cuts. However, using the pattern column generation
technique the best pattern is obtained. To get the best pattern, the knapsack sub-problem is solved first. Because
the problem is two-dimensional, there are two knapsack subproblems, namely the length and width knapsack
subproblems. Then, the dynamic programming method is used to solve them.

References
[1] P. C. Gilmore and R. E. Gomory, “A linear programming approach to the cutting-stock problem,” Oper. Res., vol. 9, pp. 849–859,
1961.
[2] W. L. Winston, Operations Research: Applications and Algorithms, 4th ed. Belmont: Thomson Brooks/ Cole, 2004.
[3] S. Wongprakornkul and P. Charnsethikul, “Solving one-dimensional cutting stock problem with discrete demands and capacitated
planning objective,” J. Math. Stat., vol. 6, pp. 79–83, 2010.
[4] F. Furini, E. Malaguti, R. Medina Durán, A. Persiani, and P. Toth, “A column generation heuristic for the two-dimensional two-
staged guillotine cutting stock problem with multiple stock size,” Eur. J. Oper. Res., vol. 218, pp. 251–260, 2012.
[5] S. Juttijudata and P. Sudjarittham, “Two-dimensional cutting stock problems with a modified column generation method,” Curr.
Appl. Sci. Technol., vol. 20, pp. 217–225, 2020.
[6] A. Mellouli, F. Masmoudi, I. Kacem, and M. Haddar, “A hgybrid genetic algorithm for optimization of two-dimensional cutting-
stock problem,” Int. J. Appl. Metaheuristic Comput., vol. 1, pp. 34–49, 2010.
[7] K. Novianingsih, R. Hadianti, and S. Uttunggadewa, “Column generation technique for solving two-dimensional cutting stock
problems: method of stripe approach,” J. Indones. Math. Soc, vol. 13, pp. 161–172, 2007.
[8] X. Song and J. A. Bennell, “Column generation and sequential heuristic procedure for solving an irregular shape cutting stock
problem,” J. Oper. Res. Soc., vol. 65, pp. 1037–1052, 2017.
[9] S. Sumetthapiwat and B. Intiyot, “A column generation on two-dimensional cutting stock problem with fixed-size usable leftover
and multiple stock sizes,” Int. J. Logist. Syst. Manag. Manag., vol. 35, pp. 273–288, 2020.
[10] P. V. Carvajal, G. Camacho, D. C. Usaquen, and D. A. Martinez, “Solving the two-dimensional knapsack problem considering
cutting-time and emission of particulate matter in the metalworking industry,” IEEE Lat. Am. Trans., vol. 16, pp. 2888–2895,
2018.
[11] M. Dolatabadi, A. Lodi, and M. Monaci, “Exact algorithms for the two-dimensional guillotine knapsack,” Comput. Oper. Res.,
vol. 39, pp. 48–53, 2012, doi: 018.
[12] M. A. Hatefi, “Developing column generation approach to solve the rectangular two-dimensional single knapsack problem,” Sci.
Iran., vol. 24, pp. 3287–3296, 2017.
[13] K. Kim, B. Kim, and H. Cho, “Multiple-choice knapsack-based heuristic algorithm for the two-stage two-dimensional cutting
stock problem in the paper industry,” Int. J. Prod. Res., vol. 52, pp. 5675–5689, 2014.
[14] K. J. Woo, “Application of fitness switching genetic algorithm for solving multidimensional knapsack problem,” J. Theor. Appl.
Inf. Technol., vol. 97, pp. 4837–4849, 2019.
[15] H. A. Taha, Operations Research: An Introduction, 8th ed. Newark: Person Education Hall, 2007.

Abdul Rahman. et. al. "Approximation Method for Solving TwoDimensional Cutting-Stock
Problems." IOSR Journal of Mathematics (IOSR-JM), 19(1), (2023): pp. 29-34.

DOI: 10.9790/5728-1901022934 www.iosrjournals.org 34 | Page