Beruflich Dokumente
Kultur Dokumente
Abstract: The aim of this paper is to introduce a proposed framework for Hierarchical Production
Planning (HPP) in the textile industry. A three level HPP is adopted; at the first level, the
decoupling point concept is used to split the full production line into smaller production stages,
followed by solving the problem of where to hold stock to provide a high level of service. Using
Theory of Constraints (TOC), the bottleneck stage has been identified in the textile production
cycle. In the second level of the HPP, the decision of determining which products to manufacture
to-order and which products to manufacture to-stock in the bottleneck stage is done using a cluster
analysis approach and demand variability analysis, in the form of Relative Demand Volatility
(RDV). At the third level, the required demand and the available capacity at the bottleneck stage
are balanced and the decision of allocating the production orders to the planning periods is made.
Keywords: Textile industry, hierarchical production planning, customer order decoupling point,
theory of constraints, relative demand volatility, master production schedule
Introduction
According to the United Nations conference on trade and development [1], during the last two
decades textiles and clothing were the second most dynamic products in world trade. The sector is
important for developing countries, which supply about 50 % of the world market for textiles and
over 60 % of the world market for clothing. On the other hand, diversification into export categories
with greater value added than traditional agricultural exports remains a major objective for many
low-income developing countries.
A textile production unit is characterized by a multi-stage process with multiple production units per
stage (i.e., parallel machines). Textile manufacturing systems involves more than one stage with
each stage yielding a product that is either pulled as finished product or further reprocessed in the
next stage. A different production planning method may be used for each production stage are
required [2].
Fibers are the raw materials for all fabrics. Some fibers are natural that come from plants, animals,
and minerals. Other fibers are synthetic that are made from petrochemicals. From the natural
fibers in Egypt, cotton is the largest crop produced and processed in the textile industry.
Litrature Review
Bowers and Jarvis [3] proposed a hierarchical production planning and scheduling model
encompassing an apparel production planning system. Adamopoulos et al. [4] presented a
decision support system dealing with the production planning and scheduling in the textile industry.
Ford and Rager [5] described an expert system developed to support a decision scenario in the
textile industry. Karacapilidis and Pappis [2] presented an interactive model based system for the
management of production in textile production systems. Shiroma and Niemeyer [6] presented the
details of the implementation of an evolutionary algorithm that made use of domain-dependent
information to generate production schedules for a manufacturer of women's clothing. Tucci and
Rinaldi [7] presentd the experience of applying the tabu search metaheuristic to production
scheduling in weaving. De Toni and Meneghetti [8] improved production planning process in a
time-based logic for a network of firms in the textile-apparel industry. Wu and Chang [9] described
the method and procedure for optimizing a textile dyeing manufacturing. Ferraris and Morini [10]
built a simulation of production scheduling at spinning stage. Laoboonlur [11] solved the dyeing
and finishing scheduling problem by using the Virtual Factory plus family scheduling. Mohafiqul et
al. [12] dealt with a decision making of apparel export-oriented firm regarding their capacity
utilization during production. Vassileva [13] described an application of a software system at the
spinning department in a textile enterprise.
Most research in the area of production planning in textile industry concentrate on studying the
operational level and the sequencing of the jobs. The tactical level planning seldom has been
tackled as input for the studied operational level. Moreover, hybrid make-to-order and make-to-
stock policy did not take enough attention in production management literature on textile industry,
although it is an important issue in production planning and in product-mix decisions as it will be
further illustrated in this work.
The textile industry is well known for unique characteristics when compared to industries in either
process or discrete manufacturing [2]. Both discrete and continuous operations are present in
textile industry. Spinning, weaving and dyeing stage in textile are continuous operations, but, the
entire system includes both continuous operation and discontinuous operations [14]. The mixed
character of a textile production system, which lies between job-shop and flow-shop, makes
production management quite complex. Hierarchical production planning approach is best suited
for such industry. The essential idea of having such hierarchical approach is the partition of a large
problem into smaller, more manageable sub-problems.
This paper presents a HPP framework that assists the decision maker in handling the complexity
of the planning process in textile industry. Using the proposed framework, the potential stocking
points can be located and the bottleneck stage is to be identified. Production and inventory policies
will be assigned for the bottleneck stage. Finally, the bottleneck stage will be optimized to achieve
a proper inventory and production decisions.
The Proposed Framework
In order to reduce the complexity of the production planning process in textile industry, a three
levels Hierarchical Production Planning (HPP) approach is proposed, as shown in figure 1. At the
first level of the HPP, the decoupling point concept is used to split the full production line into
smaller production stages, then, the problem of where to hold stock is solved to provide a high
service level. Moreover, the Customer Order Decoupling Point (CODP) concept suggests a
qualitative way to answer the MTO/MTS question. Using the Theory of Constraints (TOC), the
bottleneck stage is identified.
At the second level of the HPP, the decision of determining which products to manufacture to-
order, and which products to manufacture to-stock in the bottleneck stage is done using a cluster
analysis approach and demand variability analysis, in the form of Relative Demand Volatility
(RDV). Using cluster analysis, the products can be categorized into four clusters:
high volume, low variability;
high volume, high variability;
low volume, low variability ; and
low volume, high variability.
Accordingly, the products can be classified into two groups: high priority products, which have
large orders, and low priority products, which have small orders. These two groups will be handled
separately in terms of responsiveness. This categorization will help to exploit the bottleneck in a
better way.
Due to the large product mix in textile industry, determining which product to make-to-order versus
make-to-stock is an important issue in the production planning and in the product-mix decision.
The objective of the product-mix decision in the overall production plan is to find the product mix
and the production program that minimizes the cost subject to constraints imposed by resource
limitations, market demand, and sales forecast (the costs are production cost, inventory holding
cost, and penalty of unsatisfied demand). Such decision implies utilizing limited resources to
maximize the net value of the output from the production facilities.
At the third level of the HPP, the required
demand and the available capacity are Strategic planning
balanced and the decisions of allocating the
production orders to the planning periods are
Splitting the full
made as follows: Process and production line,
Product and
stock Identifying Customer
Level One
The first cluster is to develop a Master characteristics Order Decoupling
market
characteristics
Points, and taking
Production Schedule (MPS) model that MTO vs. MTS decision
can be implemented in the bottleneck among stages
stage.
Identify the bottleneck at the textile
The products for the second cluster will production line using TOC
Tactical Level
sufficient safety stock should be held to
meet the customer service objective. Classifying the products into
clusters using cluster
Level Two
Since clusters three and four have low analysis technique at the
bottleneck stage
volume demand, they are more likely to
be fulfilled within the delivery due data. The
Take MTO vs. MTS decision at the
The
variability Volume
Consequently, it is reasonable to produce of the
bottleneck stage for each cluster
of the
them to-order. Moreover, the fabrics for demand demand
Fiber
Warp Starch Weaving Dying/ Sewing
making &
spinning making Finishing
Raw FG
Material Inventory Inventory Inventory Inventory Inventory Inventory
Inventory (2) (3) (4) (5) (6) (7)
(1)
Figure 2 . Potential decoupling points within the stages of cotton based textile industry
Level two: taking MTO vs. MTS decisions at the bottleneck stage
Using the Theory of Constraints (TOC), it is found that the weaving stage is a bottleneck in the
cotton based textile production cycle. Therefore, in the second level, the decision of determining
which products to manufacture to-order and which products to manufacture to-stock in the
weaving stage is done using a cluster analysis approach and the demand variability analysis in the
form of “Relative Demand Volatility” (RDV).
(1) MTO and MTS decisions at weaving stage
A demand variability analysis, in the form of “Relative Demand Volatility” (RDV), as suggested in
Olhager [15] and Van Donk et al [16], could be followed. In this approach, RDV and average
demand volume are used to categorize products into MTO and MTS. The RDV is defined as the
coefficient of variation, i.e. the ratio between standard deviation of demand and the average
demand.
RDV =
i
, where i = Standard deviation for the demand of product i
( Av.Demand ) i
Clustering is the partitioning of a data set into subsets (clusters), so that the data in each subset
share some common trait. The computational task of classifying the data set into k clusters is often
referred to as K-clustering. The K-means algorithm assigns each point to the cluster whose center
is nearest. Using the Cluster Analysis, the products would be categorized into four clusters: (1)
High volume, low variability, (2) High volume, high variability, (3) Low volume, low variability and
(4) Low volume, high variability, this categorization depends on the average monthly demand and
the coefficient of variation. Figure 3 illustrates an example of the products’ classification clusters
for a typical case.
(2) Exploitation of the bottleneck stage
In order to achieve better exploitation to the bottleneck stage, a solution is proposed for finding
suitable production policy for each cluster.
(a) Classification of the products: The fabrics are classified into two groups: high priority fabrics,
which have large orders volume, and low priority fabrics, which have small orders volume. The due
date of high priority fabrics should be treated more strictly than the one of low priority fabrics.
Moreover, the high priority products customers should be provided with a high degree of
responsiveness.
(b) Fabrics with high volume and low variability: In order to facilitate fast response to the high
priority fabrics customers, the high volume items with low variability will be made-to-stock.
(c) Fabrics with high volume and high variability: If the RDV value is high, it is not reasonable to
use an MTS policy, since this would lead to excessive inventories. Therefore, the products for this
cluster will follow make-to-order policy. However, sufficient safety stock will be held to meet the
customer service objective. Safety stock (SS) is held as a hedge against uncertainty. This stock
will be used only in the case in which the delivery lead time happened to be less than the
production lead time.
(d) Fabrics with low volume and high / low variability: These two clusters have low volume
demand. Therefore, they are more likely to be fulfilled within the due date. Consequently, it is
reasonable to produce them to-order. Moreover, the fabrics of these two clusters are classified as
low priority fabrics so backorder maybe allowed.
In the next level, a master production schedule model is developed to be implemented in the
weaving stage.
X it J it Dit i t 1 i u 3, 4 (4)
X it Ctu t i u 1 (6)
1