Computers & Industrial Engineering 65 (2013) 689703

Contents lists available at SciVerse ScienceDirect

Computers & Industrial Engineering

journal homepage: www.elsevier.com/locate/caie

Dynamic Drum-Buffer-Rope approach for production planning

and control in capacitated ow-shop manufacturing systems q
Patroklos Georgiadis , Alexandra Politou
Industrial Management Division, Department of Mechanical Engineering, Aristotle University of Thessaloniki, P.O. Box 461, 541 24 Thessaloniki, Greece

a r t i c l e

i n f o

Article history:
Received 28 March 2012
Received in revised form 22 February 2013
Accepted 23 April 2013
Available online 3 May 2013
Keywords:
Drum-Buffer-Rope
Time-buffer
Flow-shop
Production planning and control
Simulation
System dynamics

a b s t r a c t
Drum-Buffer-Rope-based production planning and control (PPC) approaches provide production
managers with effective tools to manage production disruptions and improve operational performance.
The corner stone of these approaches is the proper selection of time-buffers which are considered as
exogenously dened constant. However, the majority of real-world manufacturing systems are
characterized by the dynamic change of demand and by stochastic production times. This fact calls for
a dynamic approach in supporting the decision making on time-buffer policies. To this end, we study a
capacitated, single-product, three-operation, ow-shop manufacturing system. We propose a dynamic
time-buffer control mechanism for short/medium-term PPC with adaptive response to demand changes
and robustness to sudden disturbances in both internal and external shop environment. By integrating
the control mechanism into the ow-shop system, we develop a system dynamics model to support
the decision-making on time-buffer policies. Using the model, we study the effect of policies on shop
performance by means of analysis of variance. Extensive numerical investigation reveals the insensitivity
of time-buffer policies to key factors related to demand, demand due date and operational characteristics
such as protective capacity and production times.
2013 Elsevier Ltd. All rights reserved.

1. Introduction
Insufcient production planning in manufacturing systems often
turns a non-bottleneck resource to capacity constraint resource
(CCR), which operates as a bottleneck with on average excess capacity
(Goldratt, 1988). Drum-Buffer-Rope (DBR)-based production
planning and control (PPC) approaches focus on the synchronization
of resources and material utilization in CCRs of manufacturing
systems (Goldratt & Fox, 1986; Sivasubramanian, Selladurai, &
Rajamramasamy, 2000). This synchronization calls for time-buffers
that protect the production plans of CCR from the effects of
disruptions at the preceding production resources. By means of
time-buffers (i.e. constraint, assembly, shipping time-buffers),
buffer management monitors the inventory in front of protected
resources to effectively manage and improve systems performance
(Schragenheim & Ronen, 1990; Schragenheim & Ronen, 1991).
The research agenda on the efciency of DBR approach in PPC of
manufacturing systems has received increased attention during
the last decade. The basic assumption in all relative studies is the
exogenous determination of time-buffers as a constant throughout
the planning horizon. However, the majority of real-world
q

This manuscript was processed by Area Editor Manoj Tiwari.

Corresponding author. Tel.: +30 2310 996046; fax: +30 2310 996018.
E-mail addresses: geopat@eng.auth.gr, geopat@auth.gr (P. Georgiadis), apolitou@
auth.gr (A. Politou).
0360-8352/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.cie.2013.04.013

manufacturing systems are characterized by the dynamic change

of demand and by stochastic production times. Therefore, the
decision making on time-buffer policies calls for a dynamic
mechanism. This is exactly the purpose of this paper. More
specically, we consider a dynamic, capacitated, single-product,
three-operation, ow-shop production system. We dene as
production time-buffer (PTB), the total of constraint and shipping
time-buffers. We propose a dynamic, goal-seeking, feedback
mechanism to dene PTB for short/medium-term PPC. By integrating
the proposed mechanism into the ow-shop system, we develop a
system dynamics (SD) model to support the decision making on
PTB policies. We study the shop response (dynamics of product
ows, inventories, performance measures) to PTB policies under
stochastic demand and production times. Since the dynamic
behavior may be used to evaluate the efciency of a specic PTB
policy, the SD model can be viewed as a decision support system
(DSS) for PTB-related decisions. In particular, by continuous
monitoring, the actual level of PTB is adjusted to demand-driven
desired values. The innovative element of the control mechanism
is the endogenous denition of desired PTB values. In addition,
the mechanism provides robustness to sudden disturbance occurrences in demand and shop operations. This is a positive property
to cope with uncertainty issues in both external and internal shop
environment. Using the SD model, we determine PTB increase/
decrease policies throughout a given planning horizon and we
study their effect on shop performance by means of analysis of

690

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

variance (ANOVA). The examination of results obtained by

extensive numerical investigation reveals the insensitivity of PTB
policies to key factors related to demand, demand due date and
operational characteristics such as protective capacity and production times. This is an additional appealing feature of the proposed
PTB control mechanism which provides production managers with
exibility on PTB-related decisions.
The rest of the paper is organized as follows. Section 2
presents the literature review on DBR studies and applications in
manufacturing systems and justies the suitability of SD
methodology in developing dynamic DBR-based PPC systems.
Section 3 contains the ow-shop system under study and its performance measures, the description of the SD model, the mathematical
formulation and the models validation. Section 4 presents the control parameters under study along with their sets of values, while
Section 5 presents the adaptability and robustness properties of
the dynamic PTB control mechanism. The effect of PTB policies on
the shops performance obtained by numerical investigation is given
in Section 6. Finally, in Section 7 we wrap-up with a summary, the
limitations of our work and directions for model extensions.

2. Literature review
Review papers present a variety of PPC problems dealing with
ow-shop scheduling in manufacturing systems; either with
sequence-independent set-up times or sequence-dependent
set-up times (Hejazi and Saghaan, 2005; Zhu & Wilhelm, 2006).
The proposed scheduling methods include: (i) exact methods such
as dynamic programming (Held & Karp, 1962), branch-and-bound
(Grabowski, Skubalska, & Smutnicki, 1983), integer programming
(Frieze & Yadegar, 1989) and complete enumeration; (ii) heuristic
methods such as Palmer algorithm (Palmer, 1965), Gupta
algorithm (Gupta, 1971), CDS algorithm (Campbell, Dudek, &
Smith, 1970) and NEH algorithm (Nawaz, Enscore, & Ham, 1983);
and (iii) metaheuristic methods such as simulated annealing-SA
(Liu, 1999), genetic algorithms-GA (Reeves, 1995), tabu search-TA
(Widmer & Hertz, 1989), greedy approaches (Carpov, Carlier, Nace,
& Sirdey, 2012), variable-depth search approach (Jin, Yang, & Ito,
2006), pilot methods (Vob, Fink, & Duin, 2005), hill climbing
procedures (Nearchou, 2004), ant colony system-ACS (Rajendran
& Ziegler, 2004), articial neural network-ANN (Lee & Shaw,
2000) and hybrid algorithms (Wang & Zheng, 2003).
For the specic make-to-order ow-shop environment, Stevenson,
Hendry, and Kingsman (2005) provide a detailed review on the
employed PPC approaches. The commonly used approaches include
Constant Work In Process-CONWIP (Framinan, Gonzlez, &
} rer, Stevenson,
Ruiz-Usano, 2003), Workload Control-WLC (Thu
& Silva, 2011), Material Requirement Planning-MRP (Bertrand &
Muntslag, 1993), Just-in-Time-JIT (Singh & Brar, 1992), Theory of
Constraints-TOC (Atwater, Stephens, & Chakravorty, 2004; Goldratt
& Cox, 1984; Mabin & Balderstone, 2003), Paired cell Overlapping
Loops of Cards with Authorization-POLCA (Riezebos, 2010) and
web- or e-based Supply Chain Management-SCM (Cagliano, Caniato,
& Spina, 2003; Kehoe & Boughton, 2001). The comparison of MRP,
TOC and JIT approaches justies the TOC to be more effective for a
pure ow shop or general ow shop system, when the bottleneck
resources are stationary positioned in the production process.
The effectiveness of TOC approaches is further discussed for highly
customized industries facing difculties in estimating in advance
the processing times (Stevenson et al., 2005). This is due to the fact
that TOC requires data accuracy only in CCR to control the plant
throughput (Gupta & Snyder, 2009).
TOC was rst developed in the mid-1980s (Goldratt & Cox,
1984; Gupta, 2003). It uses the DBR production scheduling
approach; production process is scheduled to run in accordance

with the needs of the CCR, as CCR determines the performance of

the whole production system. The advantages of TOC are discussed
in various industrial implementations reporting the reduction of
inventories by 49% and the improvement of due date and nancial
performance by 60% (Gupta, 2003; Mabin & Balderstone,
2003). The statistical analysis of a survey with questionnaires to
manufacturing managers performing TOC, JIT and traditional
methods provides further insights regarding the superiority of
TOC approaches on other approaches in terms of nancial and
operational performance measures (Sale & Inman, 2003). TOC has
also been used for the determination of optimal, or near optimal,
product mix decisions (Aryanezhad & Komijan, 2004; Souren,
Ahn, & Schmitz, 2005).
In certain studies, the usefulness of DBR logic in PPC of
manufacturing systems is revealed by conducted simulation experiments under different manufacturing settings. In these studies
time buffers are not optimized and remain constant throughout
the simulation process. In a make-to-stock environment,
DBR-based PPC is combined to manufacturing expediting of
products (Schragenheim, Cox, & Ronen, 1994). In a make-to order
environment, DBR approach is combined with different order
review/release policies (Russell & Fry, 1997) and it is compared
to the previously used approach in furniture manufacturing rms
(Wu, Morris, & Gordon, 1994). In a serial production line with
exponentially distributed processing times setting, DBR-based
PPC is compared to CONWIP approach (Gilland, 2002). The difference between the two approaches is that in CONWIP approach
material units are released into the line at a rate equal to line
throughput, while in DBR approach at the rate they are produced
at CCR. The outperformance of DBR approach is proved to increase
as CCR moves closer to the rst operation of production as well as
when the required throughput or service level is close to the
systems capacity (Framinan et al., 2003). Finally, in a ow-shop
setting, Sirikri and Yenradee (2006) employ DBR-based PPC and
investigate how buffer sizes related to lead time up to CCR affect
specic performance measures.
The analytical approaches to determine time-buffer sizes based
on queuing theory are limited in simple PPC manufacturing
problems. In these approaches the constraint resource is modeled
as a M/M/1/K system (Radovilsky, 1998), while the production
system is modeled either as a determination model, where a tree
structure represents the relationship between the constraint
machine and its feeder machines (Tu & Li, 1998; Ye & Han, 2008)
or as multiproduct open queuing network in which the production
operations are modeled as GI/G/m (Louw & Page, 2004).
The applicability of DBR in real world case studies is denoted by
a lot of DBR implementations in manufacturing rms; e.g. in
Orko-Pak in Netherlands that manufactures packaging material
from corrugated cardboard (Riezebos, Korte, & Land, 2003), in
Oregon Freeze Dry processing products by removing water at low
temperature and pressure (Umble, Umble, & von Deylen, 2001), in
Alameda Naval Aviation Depot that remanufactures aircraft, jet
turbine engines, engine components and avionics equipment
(Guide & Ghishelli, 1995), in a light assembly rm for heavy duty
trucks and trailers (Pegels & Watrous, 2005) and in a bearing
manufacturing company (Steele, Philipoom, Malhotra, & Fry,
2005). DBR is even used in a ghter squadron of the Israeli Air
Force for better scheduling of its missions and allocating crews
to aircraft (Ronen, Gur, & Pass, 1994).
DBR literature suggests several performance measures in
evaluating the efciency of the proposed approaches. These
measures include: the average system throughput, the average
nished product inventory, the average number of stockouts
(Duclos & Spencer, 1995); the throughput, the utilization of
machines, the average wait time of items, the percentage of
machine blocking (Mahapatra & Sahu, 2006); the mean percent

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

completion to schedule, the mean work-in-process (WIP), the

mean throughput rate, the product mean time in system (Guide,
1996); the total system output, the average and standard deviation
of ow time (Cook, 1994); and the average wait time of items, the
average WIP in queue and the system throughput (Betterton & Cox,
2009).
In the above-mentioned DBR studies, time-buffer is considered
as an exogenously dened constant throughout the planning
horizon. Further limitation is the inability to handle the non-linear,
non-stationary and uncertain nature of production process that
characterizes the majority of ow-shop manufacturing systems.
These limitations call for a dynamic time-buffer control mechanism, suitable for monitoring and adjusting time-buffer values to
desired levels. Such a control mechanism can be provided by the
usage of SD methodology. Therefore, SD is the primary modeling
and analysis tool used in this paper. Forrester (1961) introduced
SD in the early 1960s as a modeling and continuous simulation
methodology for decision-making in complex dynamic industrial
management problems (GrBler, Thun, & Milling, 2008). In
contrast to the traditional discrete event simulation-based
DSS, the methodology provides an understanding of changes
occurring within a manufacturing environment, by focusing on
the interaction between physical ows, information ows, delays
and policies that create the dynamics of the variables of interest
and thereafter searches for policies to improve system performance (Georgiadis & Michaloudis, 2012; Sterman, 2000). The
structure of a SD model is described by stocks and ows. This
structure provides the PPC with a capability to capture the
dynamics of material, product and information ows under causal
effects originated from the internal and external shop environment. The SD discipline acknowledges at the outset that realistic
representations may include non-linear elements, so closed-form
solutions are bypassed in favor of a simulation methodology.
Although the rich body of SD studies in PPC issues, DBR-based
approaches are very limited. In particular, these studies introduce
the potential use of SD theory along with the expected advantages.
Wixson and Mills (2003) present a numerical example of DBR
production process and show how SD may help in understanding
the system constraints. They consider the time-buffer of DBR as
an external parameter and assume constant production times
and innite raw materials. The potential use of SD in developing
DBR-based PPC systems for a make-to-order, ow-shop system is
presented by Politou and Georgiadis (2008). They assume
exponentially distributed production times, nite raw materials
and constant PTB throughout the planning horizon. This paper
takes the last research further by developing an endogenous
control mechanism for PTB and integrating into a SD-based PPC
system for ow-shop operations. The new possibilities provided
by the proposed SD model can be summarized in its ability to cope
with the challenges for dynamic DBR approaches in a stochastic
ow-shop environment. This ability contributes to the PPC
function of ow-shop manufacturing systems providing stable,
controllable and adaptive production plans which deal with the
non-linear, non-stationary and uncertain nature of production
processes.

3. The SD model
3.1. The ow-shop system under study
We consider a three-operation, capacitated ow-shop that
produces a single product and purchases one type of raw material
(referred as material at the remainder of paper). The CCR of the
ow-shop lies in its second operation and production rates are
dened by DBR approach. The demand follows a normal

691

distribution. Each operation of the ow-shop is considered as a

queueing model M/M/1, in which product arrival is described by a
Poisson process with the parameter k being equal to mean demand.
The capacities of the three operations are described by a Poisson
process and their mean values are set equal to Capi_M (i = 1, 2, 3).
The mean value of the production time at each operation (1/li) is
dened in Eq. (1) by means of the respective mean capacity:

li

1
Capi

1
M

Therefore, production times follow exponential distribution

with parameters li equal to or greater than k. Consequently, the
shop operation is considered as a series of three queuing models
M/M/1 (Hillier & Lieberman, 1995). Thus, the mean production
time (MPT) which is the mean value of the total production time
of shops operations is given in Eq. (2):

MPT

3
X
1

li
i1

3
X
i1

1
Capi

2
M

The notation and the respective units of measure are given in

Table 1.

Table 1
Notation list and units of measure.
Flow-shop variables
Capi
Capi_M
CapCCR
CapCCR,m
D
Dm
DSD
DB
DBDR
DBIR
DDD
ED
FPI
MCR
MF
MFO
MFODR
MFOIR
MI
MLT
MO
MOR
MP
MPR
MPT
MRR
MRT
MUR
OB
OR
ORR
ORT
PDF
PORR
PRi
PTB
r
SR
WIP0
WIPi
1/c
1/li

Capacity of the i operation (i = 1, 2, 3), items/day

Mean value of Capi (i = 1, 2, 3), items/day
Capacity of the CCR operation, which is equal to Cap2, items/day
Mean of CapCCR, items/day
Demand, items/day
Mean value of D, items/day
Standard Deviation of D, items/day
Demand Backlog, items
Demand Backlog Decrease Rate, items/day
Demand Backlog Increase Rate, items/day
Demand Due Date, days
Expected Demand, items/day
Finished Product Inventory, items
Material Consumption Rate, kg/day
Material Factor (i.e. for production of one item, MF kg of raw
material are required), kg/item
Material For Order, kg
Material For Order Decrease Rate, kg/day
Material For Order Increase Rate, kg/day
Material Inventory, kg
Material Lead Time, days
Material Order, kg
Material Order Rope, items/day
Material Procurement, kg
Material Procurement Rate, kg/day
Mean Production Time, days
Material Release Rate, kg/day
Material Release Time, days
Material Usage Rate, kg/day
Orders Backlog, items
Order Release backlog, items
Order Release Rate, items/day
Orders Rate, items/day
Planned Demand Fulllment, items
Planned Order Release Rate, items/day
Production Rate of the i operation (i = 1, 2, 3), items/day
Production Time-Buffer, days
Delay time used in computation of MOR, days
Shipments Rate, items/day
MI that has been released in production process and waits to be
processed at rst operation, kg
Work-In-Process inventory at the i operation (i = 1, 2), items
Smoothing factor used in computation of ED, 1/days
Mean value of production time of the i operation (i = 1, 2, 3),
days/item

692

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

3.2. Performance measures

The efciency of PTB policies in association with the
manufacturing process is obtained at the end of a given planning
horizon using performance measures suggested by the DBR
literature. Besides, the efciency of PTB policies in association with
the evaluation process is obtained at the end of the planning
horizon using two performance measures related to PTB. All these
measures are shown in Table 2.
3.3. Conceptual modeling
In SD discipline, causal-loop diagrams are maps of the systems
under study showing the causal links among the incorporated variables (Sterman, 2000). The generic causal-loop diagram of the
developed SD model is depicted in Fig. 1.
The control mechanisms shown in Fig. 1 are presented in the
following subsections. To improve appearance and distinction
among the variables in the causal-loop diagrams, we changed the
letter style according to the variable style; stock (state) variables
are written in capital letters, ow variables in small plain letters
and auxiliary variables in small italic letters. Stocks integrate their
ows, characterize the state of the system, give systems inertia
and provide it with memory. The arrows (inuence lines) represent
the relations among variables. The direction of the inuence lines

Table 2
Performance measures.
Performance measures of manufacturing process
ARM
Average value of material inventory, kg
AWIP1 Average value of WIP1, items
AWIP2 Average value of WIP2, items
AFP
Average value of nished product inventory, items
ADB
Average value of demand backlog, items
DBD
Demand Backlog Delay; i.e. total time duration of demand backlog
occurrence, days
ALT
Average lead time, days
APR
Average value of CCR production rate, items/day
PI
Production Index measuring the efciency of DBR-based PPC
approach; i.e. the average value of the ratio of actual CCR production
rate values over the magnitude assuming innite inventory WIP1,
dimensionless
Performance measures of PTB evaluation process
APTB
Average value of PTB, days
PTB
PTB at the end of the simulation process, days

displays the direction of the effect. The sign + or  on each

inuence line exhibits the sign of the effect. A + () sign
signies that the variables change in the same (opposite) direction.

3.3.1. Material release rate control mechanism

Fig. 2 depicts the material release rate control mechanism (for
the notation the reader must refer to Table 1). Since the corner
stone of DBR systems is the synchronization of resources and
material utilization, the material release schedule is the driving
force for the production planning. Material release (Material
Release Rate, MRR) is based on order release (Order Release Rate,
ORR). More specically, the plan of order release (Planned Order
Release Rate, PORR) is set by means of a pipeline delay of Demand
(D) with a duration equal to Material Release Time (MRT). PORR
increases ORDER RELEASE (OR) backlog, which is depleted by ORR.
ORR is limited by MATERIAL INVENTORY (MI). MRT is dened by
means of Demand Due Date (DDD), Mean Production Time (MPT),
D and PTB. The denition of MRT is based on the scheduling
logic of backward innite loading (Park & Bobrowski, 1989;
Sabuncuoglu & Karapinar, 1999); the MRT value is predicted by
back scheduling from DDD by means of MPT and PTB, in order D
to be fullled on time. The related equations are given in Appendix
A (Eqs. (A.1)(A.6)).

3.3.2. Material procurement control mechanism

In Fig. 3, the Loop 1 describes the DBR logic for the material
procurement control mechanism. More specically, Material Order
Rope (MOR) is the rope of the DBR logic and it is dened by
assuming that the material inventory is monitored for order at
the rate at which material is released in the production process
of the shop. Material For Order Increase Rate (MFOIR) increases
MATERIAL FOR ORDER (MFO), which is depleted by Material For
Order Decrease Rate (MFODR). Material Order (MO) is dened by
means of MFO. Material Procurement (MP) is the pipeline delay of
Material Order (MO) with duration equals to Material Lead Time
(MLT). Material Procurement Rate (MPR) increases MATERIAL
INVENTORY (MI), which is depleted by Material Usage Rate (MUR).
The Material procurement control mechanism refers to a review
system applying the DBR logic. In particular, this system is a periodic order quantity review system with probabilistic demand and
variable order quantity that equals MFO. The order quantity is
based on a lot-for-lot approach (Silver, Pyke, & Peterson, 1998;
Steele et al., 2005); it is equal to the material consumption of the

Fig. 1. Generic causal-loop diagram of the SD model.

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

693

Fig. 2. Causal-loop diagram of the material release rate control mechanism.

Fig. 3. Causal-loop diagram of the material procurement control mechanism.

previous period. The related equations are given in Appendix A

(Eqs. (A.7)(A.15)).
3.3.3. PTB control mechanism
The PTB control mechanism is illustrated in Fig. 4. More
specically, by monitoring PTB values a decision is continuously
made whether or not to increase or decrease its level and to what
extent. The values of PTB increase (PTB Increase) and decrease (PTB
Decrease) depend on the discrepancy (PTB Discrepancy) between
the Desired PTB and the actual level of PTB. Desired PTB is based
on DEMAND BACKLOG (DB), DDD, MPT and Expected Demand (ED),
which is a forecasted value based on the time series of D. The
magnitude of each increase or decrease is proportional to the PTB
Discrepancy at the specic time. Specically, PTB Discrepancy is
multiplied by parameters K1 for increase and K2 for decrease,
which characterize alternative PTB planning policies. Values of K1
or K2 equal to 1, in particular, refer to PTB planning policies
characterized by matching-time responsiveness. In such policies,

practically, PTB reaches Desired PTB in one time unit. Values of K1

or K2 greater than 1 refer to PTB planning policies characterized
by high-time responsiveness. In such policies, PTB reaches Desired
PTB in less than one time unit. Values of K1 or K2 smaller
than 1 refer to PTB planning policies characterized by low-time
responsiveness (i.e. PTB reaches Desired PTB in more time units).
PTB control mechanism is based on the stock management
structure suggested by Sterman (1989). Because of its central role
in the model, the related equations are given in Section 3.4.
3.3.4. Flow-shop production control mechanism
The ow-shop production control mechanism is depicted in
Fig. 5. Production Rate 1 (PR1) is controlled by Capacity 1 (Cap1),
WIP0 and Material Factor (MF), whereas Production Rate 2 (PR2) is
limited by Capacity 2 (Cap2) and WIP1. Production Rate 3 (PR3) is
limited by Capacity 3 (Cap3) and WIP2. PR3 increases FINISHED
PRODUCT INVENTORY (FPI). The control mechanism is based on
limitation functions considering the capacity and inventory

Fig. 4. Causal-loop diagram of the PTB control mechanism.

694

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

Fig. 5. Causal-loop diagram of the ow-shop production control mechanism.

constraints (Georgiadis & Michaloudis, 2012). The related

equations are given in Appendix A (Eqs. (A.16)(A.22)).

can occur continuously. The general mathematical representation

of stocks and ows is given by the following equations:

3.3.5. Shipments rate control mechanism

The shipments rate control mechanism is depicted in Fig. 6.
Loop 2 (DB Increase Rate (DBIR), DB, Shipments Rate (SR), DB Increase
Rate (DBIR)) controls the SR of the demand fulllment process.
Planned Demand Fulllment (PDF) is the pipeline delay of D with
duration equals to DDD. DBIR increases DB, which is depleted by
DB Decrease Rate (DBDR). In case of product SR is less than PDF,
DBIR gets a positive value. However, in case of delayed fulllment
of D, SR is greater than PDF and thus DBDR gets a positive value. SR
decreases ORDER BACKLOG (OB), which is increased by D. The
control mechanism is based on the assumption that all the demand
is satised, even with delay. The related equations are given in
Appendix A (Eqs. (A.23)(A.29)).

Stockt

Inflowt  Outflowtdt Stockt 0

t0

Inflowt f Stockt; Et; P;

Outflowt gStockt; Et; P 4

where E(t) any exogenous variable and P system parameters.

The generic stock-ow diagram of the developed SD model is
given in Fig. 7. The SD model consists of two modules: ow-shop
main module; and performance evaluation module which provides
the performance dynamics.
Below we provide the mathematical formulation for the PTB
control mechanism:

PTBt

PTB Increaset  PTB Decreasetdt PTBt 0 ;

t0

DDD
2

3.4. Mathematical formulation

PTBt 0

In SD discipline the development of the mathematical model is

usually presented as a stock-ow diagram that captures the model
structure and the interrelationships among the variables (Sterman,
2000). The stock-ow diagram is translated to a system of differential equations, which is then solved via simulation. High-level
graphical simulation programs support such an analysis. The
embedded mathematical equations are divided into two main
categories: the stock (state) equations, relating the accumulations
within the system of the net ow rates, and the rate equations,
dening the ows among the stocks as functions of time. In SD
models, the stock and ow perspective represents time as
unfolding continuously; events can happen at any time; change

EDt EDt  dt

1
 Dt  EDt  dt
c
Desired PTBt minDDD; MPT  EDtdt DBt

5
6
7

PTB Increaset maxK1  PTB Discrepancyt; 0=dt

PTB Decreaset maxK2  PTB Discrepancyt; 0=dt

PTB Discrepancyt Desired PTBt  PTBt

10

By Eq. (6), ED is a rst-order exponential smoothing of D at time t,

with smoothing factor 1/c. Desired PTB is dened in Eq. (7) by means
of DDD, MPT, ED and DB. PTB increase rate (PTB Increase) and PTB
decrease rate (PTB Decrease) are the increase and decrease decisions
per day. By Eq. (8) (by Eq. (9)), PTB Increase (PTB Decrease) is

Fig. 6. Causal-loop diagram of the shipments rate control mechanism.

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

695

Fig. 7. Generic stock-ow diagram of the SD model.

proportional to the positive (negative) part of PTB Discrepancy

between the desired and actual PTB, multiplied by K1 (K2). The
magnitude of PTB Discrepancy is given by Eq. (10).
The equations for the rest of control mechanisms shown in
Fig. 1 are given in Appendix A. The appendix provides also the performance measures. Performance measures given in Eqs. (A.30)
(A.35) are based on common measures suggested by the literature
(Betterton & Cox, 2009; Cook, 1994; Duclos & Spencer, 1995;
Guide, 1996). DBD (Demand Backlog Delay) measures the time
duration of DB occurrence (Eq. (A.36)). ALT (Average Lead Time) is
dened dividing OB (Orders Backlog) by ASR (Average Shipments
Rate) (Eq. (A.37)). PI (Production Index) is the average value of the
ratio of the CCR production rate over its respective upper limit.
The latter is obtained by assuming the inventory WIP1 as innitive
(Eq. (A.38)). Therefore, PI is a measure of how efciently the WIP1
is managed in order to maximize the CCR production rate. Finally,
APTB (Average PTB) is the average value of PTB (Eq. (A.39)) and PTB
refers to the value of PTB at the end of the simulation process (Eq.
(A.40)).
The entire mathematical model is a non-linear model of 13 state
variables, 21 ow variables and a considerable number of auxiliary
variables and constants (2 array auxiliary variables, 44 scalar
variables and 13 parameters). The model is developed in the
simulation software Powersim 2.5c.
3.5. Model validation
To build condence in the model and to check its quality, we
used tests suggested by the SD literature (Sterman, 2000). In
particular, we tested that every equation of the model is
dimensional consistent. Besides, we conducted extreme-condition
tests checking whether the model behaves realistically even under

extreme policies. For instance, we checked that if there is no

demand for products (D = 0), PR1, PR2, PR3, ED, DB and Desired
PTB equal zero; if there is no available capacity for production in
the rst operation of the shop (Cap1_M = 0), PR1, PR2, PR3
equal zero, DB equals the total amount of demand backlogged
throughout the planning horizon and Desired PTB equals its
maximum possible value (i.e. equals DDD). Integration error tests
were subsequently conducted. In our model we used the Euler numeric method since the integration method RungeKutta should
be avoided in models with random disturbances such as this one
(demand is not constant). We choose a simulation horizon of 300
working days (the rst 50 days are considered as transient period)
to be able to analyze short/medium-term decisions. Moreover, we
set the integrating time step (dt) initially at 0.25 days, signicantly
shorter than the shortest value of the models time constants and
ran the model. Then we cut the dt in half and ran the model
again. The results did not signicantly change, so we chose
dt = 0.125 days (=1 working hour).

4. Control parameters and sets of values

A complete numerical investigation of the SD models response
to PTB policies would require the systematic study of problems
with various levels of the system parameters. Such a detailed
experimental design is practically impossible because of the large
number of model parameters (in total 13). Consequently, we
concentrate on two parameters controlling the demand, one
parameter controlling the demand due date, two operational
parameters controlling the shop and two parameters controlling
the PTB. These control parameters are presented in the following
subsections. The rest of system parameters are shown in Table 3.

696

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

simulation run, whereas 1 denotes that production times follow

exponential distributions.

Table 3
System parameters remaining constant throughout the simulation process.
Parameter

Value

Unit

c
CapCCR,m
Initial value of MI
MF
MLT
r

3
10
120
2
3
timestep

days
items/day
kg
kg/item
days
days

4.1. Demand control parameters

Using the parameters a and b, we connect the mean value of
demand (Dm) to the mean value of CCR capacity (CapCCR,m) and
the standard deviation of demand (DSD), as follows:

Dm a  CapCCR;m

11

DSD b  Dm

12

The parameter b stands for the coefcient variance of demand. The

parameters a and b are examined in two levels; 0.9, 0.98 for a and 0,
0.2 for b.
4.2. Demand due date control parameter
The demand due date control parameter is the DDD. It is
examined in two levels; 2 and 4. These values of DDD are based
on model running with Cap2_M equals 10 items/day, Protective
Capacity equals 0.1 and Capacity Switch equals 1. For this case,
by means of (2), the mean value of the total production time
(MPT) equals 0.28 days/item. Additionally, if a = 0.9 (a = 0.98)
meaning that the mean value of daily demand is 9 (9.8)
items/order, MPT equals 2.52 (2.72) days/order.
4.3. Flow-shop control parameters
The ow-shop control parameters are the protective capacity
and the capacity switch. Protective Capacity (Atwater et al.,
2004; Betterton & Cox, 2009) connects the mean values of
production time in CCR operation and rest shop operations, as
follows:
Protectiv e Capacity

Mean of production time in CCR  Mean of production time in no  CCR

Mean of production time in CCR

The Protective Capacity is examined in two levels; 0.1 and 0.3.

Capacity Switch may take the values of 0 and 1; 0 denotes that
all the production times of the shop are kept constant during

4.4. Production time-buffer control parameters

The decision parameters that fully describe the PTB planning
policies are the control parameters K1 and K2 (see Eqs. (8) and
(9)). Recall from Section 3.3.3 that low-time responsive PTB
policies refer to values of K1 < 1. These policies lead to delaying
the starting of production process (see Eq. (A.4) of MRT denition
in Appendix A) and consequently to lower inventories. For this
reason, the parameter K1 is examined in a range from 0 to 1. With
regard to K2, the parameter is examined in a range from 0 to 0.25.
The shorter range of K2 compared to that of K1 is explained by the
fact that we wish to be more conservative in reducing the values of
PTB in order to keep low the possibility of DB occurrence.
5. Dynamics of PTB control mechanism and properties
By assigning specic values to the control parameters and
running the model, we obtain the dynamics of stocks and ows
throughout the planning horizon. We consider as the base case
the following set of control parameters: a = 0.9, b = 0.2, protective
capacity = 0.1, capacity switch = 1. For the base case, the dynamics
of PTB control mechanism for PTB control parameters K1 = 0.2 and
K2 = 0.15 and DDD = 2 days are shown in Fig. 8a. The case of
DDD = 4 days is shown in Fig. 9a. Figs. 8a and 9a illustrate the
decisions to increase or decrease PTB values on a daily basis, for
the given set of PTB control parameters. Figs. 8b and 9b illustrate
the dynamics of PTB under different K1 and K2 values.
5.1. Transient response and dynamic equilibrium
Fig. 10 (Fig. 11a) illustrates the response of actual level of PTB
for three different sets of PTB control parameters K1 and K2
and DDD = 2 days (DDD = 4 days). We observe that in case of
DDD = 2 days, the actual level of PTB does not reach any
equilibrium even when the simulation horizon is doubled (i.e.
600 days). This is explained by the fact that for a = 0.9, MPT equals
2.52 days/order (see Section 4.2). Given that DDD = 2 days, there is
not enough time for the production to be completed on time.
However, in case of DDD = 4 days, there is enough time for the
production to be completed on time and the actual level of PTB
reaches a dynamic equilibrium; i.e. the total increase of PTB is
balanced by its total decrease throughout the simulation horizon.
More specically, PTB reaches a dynamic equilibrium towards the
value of 0.34 days with different transient periods. The different
transient periods are shown more clearly in Fig. 11b.

Fig. 8. Dynamic behavior of PTB (base case, DDD = 2 days).

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

697

Fig. 9. Dynamic behavior of PTB (base case, DDD = 4 days).

Fig. 10. Dynamic behavior of PTB under different K1 and K2 values (base case, DDD = 2 days).

Fig. 11. Transient response and dynamic equilibrium of PTB under different K1 and K2 values (base case, DDD = 4 days).

5.2. Adaptability to demand changes

5.3. Robustness to sudden disturbances

Fig. 12 illustrates the response of the actual level of PTB for a

step increase in demand. In particular, the Figure illustrates the
results for a step increase with magnitude 1 item/day (about
11% of the mean value of the daily demand) on the 100th day.
Fig. 12a (Fig. 12b) proves that the step increase in demand for
DDD = 2 days (DDD = 4 days) results in a gradual increase of PTB,
which is balanced in a higher value compared to that without
step increase. This increase of PTB is explained by an increase
of Desired PTB due to the increase of ED, DB and D (see Eqs. (6)
and (7)).

The robustness of the PTB control mechanism to sudden

disturbances in the internal and external environment is given in
Figs. 13 and 14. Fig. 13 illustrates the response of actual level of
PTB when a CCR breakdown occurs from the 100th day up to the
110th day. It is shown that breakdown results in increase of PTB
and DB, which are counterbalanced later on, in both cases of
DDD; counterbalance is completed earlier in the case of
DDD = 4 days.
Fig. 14 illustrates the dynamics of PTB actual level in case of a
pulse increase in demand with magnitude 90 items/day (equal to

698

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

Fig. 12. Adaptability of PTB in case of a step increase in demand (base case, K1 = 0.2 and K2 = 0.15).

Fig. 13. Robustness of PTB in case of CCR breakdown (base case, K1 = 0.2 and K2 = 0.15).

Fig. 14. Robustness of PTB in case of a pulse increase in demand (base case, K1 = 0.2 and K2 = 0.15).

ten times the Dm) on the 100th day. The pulse increase in demand
results in a sharp increase of PTB and DB, that are counterbalanced
later on, in both cases of DDD; counterbalance is completed earlier
in the case of DDD = 4 days .

6. The effect of PTB policies on shops performance: Numerical

investigation and concluding discussion
In order to evaluate the systems performance in a dynamic
equilibrium condition, data is collected after the transient period
(50 days) to avoid irregularities during that period. The systems
performance, in terms of performance measures given in Table 2,
is examined under 32 combinations of demand, DDD and owshop control parameters, generating by the sets of levels given in
Sections 4.1, 4.2 and 4.3. In every experiment, two critical decisions

are made: how much material to order (MO) on a daily basis and
the adjustment of PTB on hourly basis.
At rst, we examine each of the above 32 combinations under
two sets of levels of PTB control parameters/factors K1 and K2;
K1 in levels 0.5 and 1, and K2 in levels 0.125 and 0.25. For each
combination of K1 and K2, three repeat simulation runs allow the
use of ANOVA to determine whether the PTB control parameters
affect signicantly the performance measures. Therefore, the total
number of simulation runs is 32  4  3 (=384).
The ANOVA results (P-values and Partial Eta Squared) for these
simulation runs (initial ANOVA) are presented in Table 4. Since
P-values are the lowest signicance levels to reject the null
hypothesis that the independent parameter does not affect the
indicated performance measures, P-values less than the 0.05 level
of signicance show statistical signicance. Besides, Partial Eta
Squared (PES) reects the signicance of the independent

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

Table 4
P-values of initial ANOVA for the effects of all control variables on performance measures.

Indicates minor signicant effect (P-value 60.05 and PES 6 0.5).

Indicates major signicant effect (P-value 60.05 and PES > 0.5).

699

700

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

parameter compared to the errors signicance; the higher the

values of PES, the higher is the effect of the independent parameter
to the dependent factor. Table 4 presents all the rst, second and
third-order results (P-values) of the initial ANOVA and only the
signicant higher-order results. For the signicant effects (P-value
60.05), the corresponding results are classied regarding the PES
value into two categories: (i) PES values equal or less than 0.5
denoting minor effect of the independent parameter to the
dependent factor and (ii) PES values higher than 0.5 denoting
major effect of the independent parameter to the dependent factor.
The results presented in Table 4 indicate that the demand, DDD
and ow-shop control parameters have signicant rst-order
effects on the majority of performance measures of manufacturing
and PTB evaluation processes. These rst-order effects (by means
of estimated marginal means of performance measures) are
presented in Table 5. For example, increase of the values of a and
b results in increase of AWIP1, AWIP2, ADB, DBD and ALT; increase
of protective capacity value results in decrease of DBD; increase of
DDD value results in decrease of ADB and DBD.
In addition, the results shown in Table 5 indicate the effects of K1
and K2 on the performance measures of manufacturing and PTB evaluation processes. It is noticeable that there is no signicant rst-order
effect of K2. For signicant effects, Table 5 illustrates these results
for rst-order effects of K1 and K2. Table 5 indicates that:
Parameter K1 has signicant effect on the majority of
performance measures of manufacturing process (i.e. on the
measures ARM, AWIP1, AWIP2, DBD, APR and PI)
Parameters K1 and K2 do not have signicant effect on the performance measures of PTB evaluation process. However, these
measures are inuenced by the demand, DDD and
ow-shop control parameters.

The above observations lead to the necessity of more

thoroughly examination of PTB control parameters/factors K1 and
K2. Therefore, in a second ANOVA, we examine each of the 32
combinations of demand, DDD and ow-shop control parameters
under 100 sets of levels of PTB control parameters/factors K1 and
K2 generating by the combinations of their levels: K1 from 0.1 to
1 with step of 0.1; K2 from 0.025 to 0.25 with step of 0.025. For
each combination of K1 and K2, three repeat simulation runs
allow the use of ANOVA to determine whether the PTB control
parameters affect signicantly the performance measures
and select their optimum values among the considered ones.
Therefore, the total number of simulation runs is 32  100  3
(=9,200).
The ANOVA results for each of 32 combinations of demand, DDD
and ow-shop control parameters (second ANOVA) indicate that
the effect K1  K2 on all performance measures of manufacturing
process is insignicant (for level of signicance equal to 0.05).
Consequently, there is no meaning to track a specic combination
of K1 and K2 in order to optimize the performance measures.
Although it is in principle risky to generalize on the basis of
numerical examples, the embedded PTB control mechanism leads
to the conjecture that the performance measures are indeed robust
to moderate changes of the control parameters K1 and K2 for the
studied cases of demand, DDD and ow-shop control parameters of
the shop.
It is very interesting that the insensitivity of performance
measures to changes in PTB control parameters that was identied
through the above ANOVA analysis is not coincidental, nor peculiar
to a particular combination of the demand, DDD and ow-shop
control parameters. Extensive simulation results, not shown for
brevity, reveal that the performance measures are indeed robust
to changes in PTB control parameters for a wide range of demand,

Table 5
First-order Estimated Marginal Means (EMMs) of performance measures connected to control variables (initial ANOVA).

Indicates no signicant effect.

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

DDD and ow-shop control parameters. The robustness of the

performance measures to PTB control parameters and to demand
characteristics is an extremely positive property of the proposed
methodology, since from one hand accurate forecasts of demand
are in many real-world applications difcult to obtain and on the
other hand this property provides production managers with exibility in decision making with regard to parameters K1 and K2.
Besides, as it is shown in Table 6, for all the 32 combinations,
the effect K1  K2 is insignicant on the performance measures of
PTB evaluation process (for level of signicance equal to 0.05).
The insignicance of K1  K2 effect on APTB and PTB reinforce the
evidence that the dynamic PTB control mechanism is indeed robust
to changes in PTB control parameters for a wide range of demand,
DDD and ow-shop control parameters.

7. Summary, limitations and directions for model extensions

This paper was aimed to introduce a dynamic PTB control
mechanism for DBR-based PPC of ow-shop manufacturing
systems. We developed a SD model for a three-operation, singleproduct, capacitated ow-shop system that purchases one type
of raw material. We considered a normally distributed demand
and exponentially distributed production times. Integrating the
PTB control mechanism into the SD model, we examined the shop
performance under different demand, DDD and shop settings.
We proved the insensitivity of performance measures of
manufacturing and PTB evaluation processes to changes in PTB

Table 6
Results of second ANOVA for the effects of PTB control variables on performance
measures of PTB evaluation process.
a

Prot.
cap.

Cap.
switch

DDD

APTB

PTB

P-value

PES

P-value

PES

0.91

0.24

0.75

0.26

0.77

0.26

0.48

0.29

0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9

0
0
0
0
0
0
0
0

0.1
0.1
0.1
0.1
0.3
0.3
0.3
0.3

0
0
1
1
0
0
1
1

2
4
2
4
2
4
2
4

0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9

0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2

0.1
0.1
0.1
0.1
0.3
0.3
0.3
0.3

0
0
1
1
0
0
1
1

2
4
2
4
2
4
2
4

0.52
0.25
0.27
0.66
0.95
0.79
0.49
0.88

0.29
0.31
0.31
0.27
0.23
0.26
0.29
0.24

0.37
0.42
0.34
0.80
0.46
0.82
0.33
0.85

0.30
0.30
0.30
0.26
0.29
0.25
0.30
0.25

0.98
0.98
0.98
0.98
0.98
0.98
0.98
0.98

0
0
0
0
0
0
0
0

0.1
0.1
0.1
0.1
0.3
0.3
0.3
0.3

0
0
1
1
0
0
1
1

2
4
2
4
2
4
2
4

0.94
0.49

0.23
0.29

0.37

0.30

0.44
0.46

0.29
0.29

0.64

0.27

0.98
0.98
0.98
0.98
0.98
0.98
0.98
0.98

0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2

0.1
0.1
0.1
0.1
0.3
0.3
0.3
0.3

0
0
1
1
0
0
1
1

2
4
2
4
2
4
2
4

0.55
0.71
0.51
0.39
0.98
0.76
0.79
0.73

0.28
0.27
0.29
0.30
0.21
0.26
0.26
0.26

0.96
0.82
0.66
0.16
0.40
0.72
0.92
0.38

0.22
0.25
0.27
0.33
0.30
0.27
0.24
0.30

*
Indicates that the value of performance measure obtained by all simulation runs
is the same.

701

control parameters. The results obtained by extensive numerical

investigation and ANOVA analysis revealed that the PTB control
parameters are indeed robust to changes in the values of demand,
DDD and operational parameters.
The proposed PTB control mechanism provides the following
possibilities: (i) decision making on the magnitude of PTB without
tracking a specic combination of control parameters K1 and K2
(by assuming initial values for PTB, K1 and K2, managers may
decide on PTB values which are based on the evaluation of Desired
PTB values), (ii) ability to integrate real-time disturbances
(machine failures, demand increase) into PTB-related decisions
and (iii) dynamic adaptation of PTB to changes in internal and
external shop environment. In addition, the employed control
mechanism provides the ability to consider the non-linear,
non-stationary and uncertain nature of production process that
characterizes the majority of ow-shop manufacturing systems.
Without this control mechanism, it would be impossible to get
the adaptive behavior of the manufacturing system towards
all the possible changes. However, the learning period is a
precondition for successful real-world applications. This is aligned
with the transient period occurrence throughout the simulation
process presented in this paper. After all, the control mechanism
is self-correcting; its feedback structure ensures that forecast
errors, changes in the structure of the shop environment and even
self-generated overreactions can eventually be corrected.
The proposed real-time PTB control mechanism faces
limitations. The continuous monitoring and adjustment of PTB
requires the use of real-time controllers, which in automated
production systems are integrated in their IT infrastructure.
However, in a more traditional ow shop, the controllers are
human-driven and consequently the monitoring and adjustment
of PTB practically takes place in longer periods, resulting in
suboptimal shop performance. For example, for the base case, for
DDD = 2 days and for K1 = 0.2 and K2 = 0.15, the average value of
demand backlog (ADB) equals 0.51 items, when dt equals
0.125 days. However, for dt = 0.5 days, ADB equals 19.37 items.
Therefore, in automated ow-shop manufacturing systems, the
oor manager in order to exploit the advantages of the developed
model in applying DBR-based PPC must estimate the values of
Demand Due Date and PTB control parameters and ascertain the
initial values of stock variables. In case of real-time disturbances,
the manager has to ensure the updating of the values of model
parameters.
In case of traditional ow-shop systems that apply DBR-based
PPC with human-driven controllers, the oor manager, at the
beginning of each day, must decide upon the value of PTB. In this
decision, it is suggested to follow a simple rule. In particular,
based on the PTB value of the previous day and the Desired PTB
value (see Eq. (7)), the manager has to compute the value of
PTB Discrepancy (see Eq. (10)). Then, the current value of PTB is
calculated (see Eqs. (5), (8), and (9)) by using the assumed values
of control parameters K1 and K2. Finally, the current PTB value is
used to calculate the values of Material Release Time (MRT) (see
Eq. (A.4)), Planned Order Release Rate (PORR) (see Eq. (A.2)), Order
Release Rate (ORR) (see Eq. (A.3)) and Material Release Rate (MRR)
(see Eq. (A.6)), and to schedule the raw material release in
production line.
The results presented in this paper certainly do not exhaust the
possibilities of investigating all the factors affecting the dynamic
DBR-based PPC of ow-shop manufacturing systems. For example,
it is worthwhile to study the proposed PPC system assuming multi
products under different shop settings and to integrate cost
elements into the proposed control mechanism. Finally, the
development of self-adaptive mechanisms for PTB control
parameters (K1 and K2) may have added-value in the development
of more comprehensive DBR-based PPC systems.

702

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

Appendix A. Equations of control mechanisms and performance

measures

Based on the assumption that all the demand is satised, even

with delay.

A.1. Material release rate control mechanism

Based on backward innite loading scheduling (Park and
Bobrowski, 1989; Sabuncuoglu and Karapinar, 1999).

ORt

PORRt  ORRtdt ORt 0 ;

ORt 0 0

PORRt Dt  MRT

A:25

A:2

A:26

SRt minPDFt DBt; FPIt=dt

Z t
ORTt  SRtdt OBt 0 ;
OBt

A:27

MRTt max0; DDD  MPT  Dt  dt  PTBt

A:4

1
Capi

DBt 0 0 A:24

t0

DBDRt minmaxSRt  PDFt=dt; 0; DBt=dt

A:3

i1

A:23

DBIRt maxPDFt=dt  SRt; 0

ORRt minORt=dt; MIt=MF=dt

MPT

PDFt Dt  DDD
Z t
DBIRt  DBDRtdt DBt 0 ;
DBt

A:1

t0

3
X

A.4. Shipments rate control mechanism

OBt 0 0

A:28

t0

A:5

ORTt Dt=dt

A:29

MRRt ORRt  MF

A:6
A.5. Performance measures of manufacturing process

A.2. Material procurement control mechanism

PT

Based on a periodic order quantity review system with

probabilistic demand and variable order quantity that equals
MFO (Silver et al., 1998; Steele et al., 2005).
MORt ORRt  r  MF
A:7
Z t
MFOt MFOIRt  MFODRtdt MFOt0 ; MFOt 0 0 A:8
t0

MFOIRt MORt

(
MFODRt


A:9

MFOt
;
dt

if an order is giv en at time t

0;

otherwise

MFOt; if an order is giv en at time t

0; otherwise

MOt

A:10

A:11

t1 MI t

ARM

T
PT

t1 WIP 1;t

AWIP1

T
t1 WIP 2;t
AWIP2
T
PT
t1 FPI t
AFP
T
PT
DBt
ADB t1
T
PT
PR2;t
APR t1
T

T
X
t; if DBt > 0
t t ; where t t
DBD
0; otherwise
t1

A:30
A:31

PT

OBT
; where
ASRT
PT
SRt
ASRT t1
av erage shipments rate
T
PT PR2;t

A:32
A:33
A:34
A:35
A:36

ALT

MPt MOt  MLT

Z

MIt

A:12

MPRt  MURtdt MIt0 ;

MIt 0 0

A:13

t0

PI

MPRt MPt=dt

A:14

MURt minMRRt; MIt=dt

A:15

A:38

A.6. Performance measures of PTB evaluation process

APTB

Based on limitation functions considering the capacity and

inventory constraints (Georgiadis and Michaloudis, 2012).
Z

PT

A.3. Flow-shop production control mechanism

WIP0t

t1 Cap2;t

A:37

t1 PTBt

T
PTB PTBt T see Eq: 5; in Section 3:4

A:39
A:40

MRRt  MCRtdt WIP0t 0 ; WIP0t 0 0

A:16

References

A:17

Aryanezhad, M. B., & Komijan, A. R. (2004). An improved algorithm for optimizing

product mix under the theory of constraints. International Journal of Production
Research, 42(20), 42214233.
Atwater, J. B., Stephens, A. A., & Chakravorty, S. S. (2004). Impact of scheduling free
goods on the throughput performance of a manufacturing operation.
International Journal of Production Research, 42(23), 48494869.
Bertrand, J. W. M., & Muntslag, D. R. (1993). Production control in engineer-to-order
rms. International Journal of Production Economics, 30(31), 322.
Betterton, C. E., & Cox, J. F. (2009). Espoused Drum-Buffer-Rope ow control in serial
lines: A comparative study of simulation models. International Journal of
Production Economics, 117, 6679.
Cagliano, R., Caniato, F., & Spina, G. (2003). E-business strategy: How companies are
shaping their supply chain through the Internet. International Journal of
Operations & Production Management, 23(10), 11421162.
Campbell, H. G., Dudek, R. A., & Smith, M. L. (1970). A heuristic algorithm for the njob, m-machine sequencing problem. Management Science, 16(10), B630B637.

t0

MCRt PR1t  MF
Z t
WIP1t PR1t  PR2tdt WIP1t 0 ; WIP1t 0 0
t0

PR3t min
FPIt

A:18


WIP0t
WIP1t
; Cap1t ; PR2t min
; Cap2t
PR1t min
A:19
MF  dt
dt
Z t
A:20
WIP2t PR2t  PR3tdt WIP2t 0 ; WIP2t 0 0
t0

WIP2t
; Cap3t
dt


A:21

t0

PR3t  SRtdt FPIt 0 ; FPIt 0 0

A:22

P. Georgiadis, A. Politou / Computers & Industrial Engineering 65 (2013) 689703

Carpov, S., Carlier, J., Nace, D., & Sirdey, R. (2012). Two-stage hybrid ow shop with
precedence constraints and parallel machines at second stage. Computers &
Operations Research, 39, 736745.
Cook, D. P. (1994). A simulation comparison of traditional JIT and TOC
manufacturing systems in a ow shop with bottlenecks. Production and
Inventory Management, 35(1), 7378.
Duclos, K. L., & Spencer, M. S. (1995). The impact of a constraint buffer in a ow
shop. International Journal of Production Economics, 42, 175190.
Forrester, J. W. (1961). Industrial dynamics. Cambridge, MA: MIT Press.
Framinan, J. M., Gonzlez, P. L., & Ruiz-Usano, R. (2003). The CONWIP production
control system: Review and research issues. Production Planning and Control,
14(3), 255265.
Frieze, A. M., & Yadegar, J. (1989). A new integer programming formulation for the
permutation owshop problem. European Journal of Operational Research, 40,
9098.
Georgiadis, P., & Michaloudis, C. (2012). Real-time production planning and control
system for job-shop manufacturing: A system dynamics analysis. European
Journal of Operational Research, 216, 94104.
Gilland, W. (2002). A simulation study comparing performance of CONWIP and
bottleneck-based release rules. Production Planning and Control, 13(2), 211219.
Goldratt, E. M. (1988). Computerized shop oor scheduling. International Journal of
Production Research, 26(3), 443455.
Goldratt, E. M., & Cox, J. (1984). The goal A process of ongoing improvement. North
River Press Inc.
Goldratt, E. M., & Fox, R. (1986). The race. North River Press Inc.
Grabowski, J., Skubalska, E., & Smutnicki, C. (1983). On ow shop scheduling with
release and due dates to minimize maximum lateness. Journal of the Operational
Research Society, 34(7), 615620.
GrBler, A., Thun, J. H., & Milling, P. M. (2008). System dynamics as a structural
theory in operations management. Production and Operations Management, 17,
373384.
Guide, V. D. R. Jr., (1996). Scheduling using Drum-Buffer-Rope in a remanufacturing
environment. International Journal of Production Research, 34, 10811091.
Guide, V. D. R., & Ghishelli, G. A. (1995). Implementation of Drum-Buffer-Rope at a
military rework depot engine works. Production and Inventory Management,
36(3), 7982.
Gupta, J. N. D. (1971). A functional heuristic algorithm for the owshop scheduling
problem. Operational Research Quarterly, 22(1), 3947.
Gupta, M. (2003). Constraints management recent advances and practices.
International Journal of Production Research, 41(4), 647659.
Gupta, M., & Snyder, D. (2009). Comparing TOC with MRP and JIT: A literature
review. International Journal of Production Research, 47(13), 37053739.
Hejazi, R., & Saghaan, S. (2005). Flowshop-scheduling problems with makespan
criterion: A review. International Journal of Production Research, 43(14),
28952929.
Held, M., & Karp, R. M. (1962). A dynamic programming approach to sequencing
problems. Journal of the Society for Industrial and Applied Mathematics, 10(1),
196210.
Hillier, F. S., & Lieberman, G. J. (1995). Introduction to operations research. McGrawHill.
Jin, Z., Yang, Z., & Ito, T. (2006). Metaheuristic algorithms for the multistage hybrid
owshop scheduling problem. International Journal of Production Economics,
100(2), 322334.
Kehoe, D., & Boughton, N. (2001). Internet based supply chain management: A
classication of approaches to manufacturing planning and control.
International Journal of Operations & Production Management, 21(4), 516524.
Lee, I., & Shaw, M. J. (2000). A neural-net approach to real time ow-shop
sequencing. Computers and Industrial Engineering, 38, 125147.
Liu, J. (1999). The impact of neighbourhood size on the process of simulated
annealing: Computational experiments on the owshop-scheduling problem.
Computers & Industrial Engineering, 37, 285288.
Louw, L., & Page, D. C. (2004). Queuing network analysis approach for estimating the
sizes of the time buffer in theory of constraints-controlled production system.
International Journal of Production Research, 42(6), 12071226.
Mabin, V. J., & Balderstone, S. J. (2003). The performance of the theory of constraints
methodology Analysis and discussion of successful TOC applications.
International Journal of Operations and Production Management, 23(6), 568595.
Mahapatra, S. S., & Sahu, A. (2006). Application of theory of constraints on
scheduling of Drum-Buffer-Rope. In International conference on global
manufacturing and innovation, 2729 July 2006, Rourkela, India.
Nawaz, M., Enscore, E., Jr, & Ham, I. (1983). A heuristic algorithm for the m-machine,
n-job owshop sequencing problem. OMEGA International Journal Management
Science, 11, 9195.
Nearchou, A. C. (2004). A novel metaheuristic approach for the ow shop scheduling
problem. Engineering Applications of Articial Intelligence, 17, 289300.
Palmer, D. S. (1965). Sequencing jobs through a multi-stage process in the
minimum total time A quick method of obtaining a near optimum.
Operational Research Quarterly, 16(1), 101107.
Park, P. S., & Bobrowski, P. M. (1989). Job release and labor exibility in a dual
resource constrained job shop. Journal of Operations Management, 8(3), 230249.
Pegels, C. C., & Watrous, C. (2005). Application of the theory of constraints to a
bottleneck operation in a manufacturing plant. Journal of Manufacturing
Technology Management, 16(3), 302311.
Politou, A., & Georgiadis, P. (2008). Production planning and control in ow shop
operations using drum buffer rope methodology: A system dynamics approach.

703

In 26th International conference of the system dynamics society, 2024 July 2008
Athens, Greece.
Radovilsky, Z. D. (1998). A quantitative approach to estimate the size of the time
buffer in the theory of constraints. International Journal of Production Economics,
55, 113119.
Rajendran, C., & Ziegler, H. (2004). Ant-colony algorithms for permutation owshop
scheduling to minimize makespan/total owtime of jobs. European Journal of
Operational Research, 155, 426438.
Reeves, C. (1995). A genetic algorithm for owshop sequencing. Computers and
Operations Research, 22, 513.
Riezebos, J. (2010). Design a POLCA material control system. International Journal of
Production Research, 48(5), 14551477.
Riezebos, J., Korte, G. J., & Land, M. J. (2003). Improving a practical DBR buffering
approach using workload control. International Journal of Production Research,
41(4), 699712.
Ronen, B., Gur, R., & Pass, S. (1994). Focused management in military organizations:
An avenue for future industrial engineering. Computers & Industrial Engineering,
27(14), 543544.
Russell, G. R., & Fry, T. D. (1997). Order review/release and lot splitting in DrumBuffer-Rope. International Journal of Production Research, 35(3), 827845.
Sabuncuoglu, I., & Karapinar, H. Y. (1999). Analysis of order review/release problems
in production systems. International Journal of Production Economics, 62,
259279.
Sale, M. L., & Inman, R. A. (2003). Survey-based comparison of performance and
change in performance of rms using traditional manufacturing, JIT and TOC.
International Journal of Production Research, 41(4), 829844.
Schragenheim, E., Cox, J., & Ronen, B. (1994). Process ow industry-scheduling and
control using theory of constraints. International Journal of Production Research,
32(8), 18671877.
Schragenheim, E., & Ronen, B. (1990). Drum-Buffer shop oor control. Production
and Inventory Management Journal, 31(3), 1823.
Schragenheim, E., & Ronen, B. (1991). Buffer management: A diagnostic tool for
production control. Production and Inventory Management, 32(2), 7479.
Silver, E. A., Pyke, D. F., & Peterson, R. (1998). Inventory management and production
planning and scheduling (3rd ed.). USA: John Wiley & Sons.
Singh, N., & Brar, J. K. (1992). Modelling and analysis of just-in-time manufacturing
systems: A review. International Journal of Operations & Production Management,
12(2), 314.
Sirikri, V., & Yenradee, P. (2006). Modied Drum-Buffer-Rope scheduling
mechanism for a non-identical parallel machine ow shop with processingtime variation. International Journal of Production Research, 44(17), 35093531.
Sivasubramanian, R., Selladurai, V., & Rajamramasamy, N. (2000). The effect of
the Drum-Buffer-Rope (DBR) approach on the performance of a synchronous
manufacturing system (SMS). Production Planning and Control, 11(8),
820824.
Souren, R., Ahn, H., & Schmitz, C. (2005). Optimal product mix decisions based on
the theory of constraints? Exposing rarely emphasized premises of throughput
accounting. International Journal of Production Research, 43(2), 361374.
Steele, D. C., Philipoom, P. R., Malhotra, M. K., & Fry, T. D. (2005). Comparisons
between Drum-Buffer-Rope and material requirements planning: A case study.
International Journal of Production Research, 43(15), 31813208.
Sterman, J. D. (1989). Modeling managerial behavior: Misperceptions of feedback
in a dynamic decision making experiment. Management Science, 35(3),
321339.
Sterman, J. D. (2000). Business dynamics: Systems thinking and modeling for a complex
world. New York: McGraw-Hill.
Stevenson, M., Hendry, L. C., & Kingsman, B. G. (2005). A review of production
planning and control: The applicability of key concepts to the make-to-order
industry. International Journal of Production Research, 43(5), 869898.
} rer, M., Stevenson, M., & Silva, C. (2011). Three decades of workload control
Thu
research: A systematic review of the literature. International Journal of
Production Research, 49(23), 69056935.
Tu, Y. M., & Li, R. K. (1998). Constraint time buffer determination model.
International Journal of Production Research, 36(4), 10911103.
Umble, M., Umble, E., & von Deylen, L. (2001). Integrating enterprise resources
planning and theory of constraints: A case study. Production and Inventory
Management Journal, 42(2), 4348.
Vob, S., Fink, A., & Duin, C. (2005). Looking ahead with the pilot method. Annals of
Operations Research, 136, 285302.
Wang, L., & Zheng, D.-Z. (2003). An effective hybrid heuristic for ow shop
scheduling. International Journal of Advanced Manufacturing Technology, 21,
3844.
Widmer, M., & Hertz, A. (1989). A new heuristic method for the ow shop
sequencing problem. European Journal of Operational Research, 41, 186193.
Wixson, J. R., & Mills, J. I. (2003). A system dynamics view of the theory of
constraints. In 21st International conference of the system dynamics society, 2025
July 2003, New York.
Wu, S., Morris, J., & Gordon, T. (1994). A simulation analysis of the effectiveness of
Drum-Buffer-Rope scheduling in furniture manufacturing. Computers &
Industrial Engineering, 26(4), 757764.
Ye, T., & Han, W. (2008). Determination of buffer sizes for Drum-Buffer-Rope (DBR)controlled production systems. International Journal of Production Research,
46(10), 28272844.
Zhu, X., & Wilhelm, W. E. (2006). Scheduling and lot sizing with sequencedependent setup: A literature review. IIE Transactions, 38, 9871007.