Page 20 - IJOCTA-15-1

An International Journal of Optimization and Control: Theories & Applications
ISSN: 2146-0957 eISSN: 2146-5703
Vol.15, No.1, pp.14-24 (2025)
https://doi.org/10.36922/ijocta.1704

RESEARCH ARTICLE

Significance of stochastic programming in addressing production
planning under uncertain demand in the metal industry sector

1
Seyda Karahan Orak , Nezir Aydin 1,2* , Ecem Karatas 1
1 Department of Industrial Engineering, Yildiz Technical University, Istanbul, T¨urkiye
2 College of Science and Engineering, Hamad Bin Khalifa University, Doha, Qatar
seydakrhn@gmail.com, naydin@hbku.edu.qa, karatass.ecem@gmail.com

ARTICLE INFO ABSTRACT

Article History: One of the most important disciplines for businesses is production planning.
Received 13 October 2024 Production planning involves various cost elements such as labor, equipment,
Accepted 11 December 2024 raw materials, and inventory while significantly impacting strategic aspects
Available Online 20 January 2025 like sales, profit, and market share. Mathematical models used in production
planning often address problems of cost minimization or profit maximization.
Keywords:
However, besides deterministic-based linear programming applications, it is
Stochastic programming
known that the effect of randomness also plays a significant role in production
Deterministic optimization
planning. When parameters are stochastic, meaning random, mathematical
Production planning
models must be capable of generating solutions under the influence of these
Metal industry
random parameters. Stochastic modeling developed for problems affected by
Sample average approximation
random parameters can yield the desired results. This study addresses the
AMS Classification 2010: issue of production planning using stochastic modeling for a company that
90B30; 90C05; 90C11; 90C15; 90C90 manufactures industrial-type pipe clamps and has two main product groups.
The model that minimizes costs under demand uncertainty uses the Sample
Average Approximation (SAA) approach. Initially, a deterministic model was
established to obtain the solution when randomness was not included. Sub-
sequently, the stochastic model was solved by creating different scenario sets
using SAA, and comparison results were presented.

1. Introduction all parameters are random variables are called stochastic
programming problems. 3
Since the dawn of humanity, production has been a ne- Inventory is the quantity of products an entrepreneur
cessity. Many factors influence the stages of production. holds to meet future demand. The main goal of inven-
4
Each production process creates its constraints and out- tory control is to determine the optimal stock level at the
comes, necessitating production planning. Determining right time, minimize costs for the firm, and maintain this
where, when, how, and how much to produce and how to orderly. The method used for inventory planning greatly
5
ensure production meets demand are central to production depends on the nature of the demand. Static (fixed) de-
planning. In modern businesses producing multiple prod- mand problems indicate a stable demand pattern, while
ucts, production planning is conducted to determine the dynamic (variable) demand problems show that demand
optimal quantities of produced goods. 1 changes over time. If a company knows with certainty the
6
demand quantities and times it will face in the future at
Thus, optimizing production is a fundamental way to in- the beginning of the planning horizon, it is a deterministic
7
crease efficiency. Using linear programming, a production demand. The complexity or simplicity of the inventory
planning model can be created by looking at past pro- policy created can vary based on whether the demand
duction and demand, and this model can be analyzed to is deterministic (known for sure) or stochastic (known
make predictions. One of the most important assump- probabilistically). 8
tions of linear programming is that model parameters are
2
deterministic. However, in real life problems, some param- Correct management of inventories, which have a signif-
eters may have uncertain values. Problems in which one or icant share of total costs for businesses, has become an
*Corresponding Author
14