Download Marketing-production interface through an integrated

Marketing-production interface through an integrated DSS
Ana Respício*
Maria Eugénia Captivo**
Centro de Investigação Operacional and Faculdade de Ciências da Universidade de Lisboa
Lisboa, Portugal
*Email: [email protected]
**Email: [email protected]
Abstract
This paper reports on a real-world application addressing the coordination between cross functionality of
marketing and production. We focus on the critical issues of this interface, namely: capacity planning and
long-range to medium range sales forecasting, production short-term scheduling and short-range sales
forecasting. We present a case study concerning the development of a decision support system for the paper
industry. It assists both marketing and production decision makers, leading to a reduction of conflicts. The
decision-making process is generalized into a framework that allows the sharing of the same decision models
but with different perspectives. This framework may be adapted to other industries.
Keywords
Interfaces for Marketing-Production Decision-Making, Production Planning and Scheduling, Decision Support
Systems.
1. INTRODUCTION
Taking competitive advantage is a key issue for most companies. To win and keep buyers, the companies
behavior is becoming more and more customer-oriented. Customers require short lead times at the least prices.
In addition, more flexibility is offered in the order management. Customers are allowed to cancel and make
changes in their orders. This gives instability to the industrial processes, in particular for non make-to-stock
companies, for whom the planning functions depend on knowing the demand. The study of Amaro et al. (1999)
points that, for non make-to-stock companies, customization is only a qualifier to compete with similar
companies. Besides, they also suggest that winning orders depend on the price and on the quality of customer
service, as for the make-to-stock companies. In consequence, the companies are forced to improve order
management and production planning/scheduling functions. Typically, these functions are assigned to distinct
units: the marketing and the production, respectively. Cooperation and functional coordination of these units
are fundamental aspects for a successful manufacturing system.
This work describes how a Decision Support System (DSS) for production planning and scheduling (Respício
et al. 2002) evolved to an interface between marketing and production, so reducing cross-functional conflicts. It
results from an R&D experience, which lasted for more then ten years, in cooperation with Portucel, the major
Portuguese paper producer and a leader company in the European paper market.
Our approach stresses the coordination between marketing and production decision-making. Order acceptance,
a function of the marketing, is performed using the same model that production uses to plan the capacity
allocation to production orders. That feature assures the consistency of the solution schedules. Another original
feature, we propose, is the integration of the short-term scheduling decisions in the capacity planning
algorithms. This allows for improving the capacity exploitation. Besides, a module to assist forecasting demand
is an important component. Marketing uses this module to analyze demand and perform medium-term
forecasting, while production uses it to decide about anticipating short-term demand, thus achieving economies
of scale.
Some other approaches for automated scheduling in the paper industry have been proposed. A DSS developed
for Naheola Mill is presented in (Pickard and Yeager 1997). (Keskinocak et al. 1998) and (Murthy et al. 1999)
describe an agent-based DSS that follows a multi-criteria optimization approach. The problems tackled include
allocation of orders to paper machines, sequencing of orders in each machine, scheduling cutting decisions and
vehicle loading. However, these works do not address the marketing/production interface.
In the next section, we discuss the cross functionality of marketing and production. Section 3 is devoted to the
presentation of the case study that supports this research and gives an overview of the DSS we developed. We
683
Marketing-production interface through an integrated DSS
give a brief description of the global problem, of the DSS features, and also of some aspects of the DSS
architecture and implementation. In section 4, we propose a decision-making framework that allows
coordinating marketing and production. Section 5 concludes by stressing the results achieved.
2. MARKETING/PRODUCTION CROSS FUNCTIONALITY
Marketing and production functions are organized as separate units on most of companies. Although strongly
inter-related, they are evaluated by different criteria. Marketing is concerned with revenue maximization
through sales, while production strives to minimize production costs. Marketing sets prices and advertising
policy: faced with the quantities that the customers demand, it negotiates both prices and delivery dates.
Production is then required to produce the demand at minimum cost. However, this decomposition results in
conflicts due to inconsistent objectives or externalities imposed by one function on the other, and such
decomposition may also yield suboptimal overall performance (Eliashberg and Steinberg 1993).
Literature on production planning and scheduling reflects this separate decision-making. Several authors have
suggested to combine both decision problems and to solve them jointly (Eliashberg and Steinberg 1993).
However, these approaches assume an existing central authority that can make all the decisions. For most of
companies, this doesn’t correspond to the existing functional structure. On the other hand, the complexity of
the decision processes involved requires a high level of expertise.
Production scheduling arises as an area of conflict between marketing and production. Marketing is responsible
for order acceptance and negotiates delivery dates and prices with customers. Marketing complaints about long
leading times forgetting that production capacity is limited. In the presence of unrealistic customer
commitments, the production is not able to fulfill promised delivery dates. In (Crittenden 1992), the capacity
planning problem is considered as the most critical decision problem that both units need to tackle. Several
rules are suggested, namely, rules that consider customer priorities because they allow for objective and
subjective allocation of capacity to production of orders. If we exclude pricing decisions, then planning
horizons and capacity planning are the principal components of joint marketing production decision-making,
whereas forecasting is the most important component of all the marketing functions.
To reduce conflict it is necessary to understand which variables are the sources of it and how they have arisen.
That understanding can help us direct the intervention needed.
The existing production capacity sets up the upper limit to the quantities that may be produced. Sometimes,
marketers forget that fact and accept all the incoming orders (Massey and Dawes 2000). For companies that
can easily change the production capacity, this problem is only significant in the short run. However, if the
capacity is fixed it will result in unfeasibility. Besides, marketers expect a high degree of flexibility from
production. They are aware of the importance of meeting delivery dates. Nevertheless, some times they promise
unrealistic short dates to win a sale. If the production can’t satisfy these dates, the situation is analyzed and the
order might be delivered late or the delivery date is renegotiated with the customer. In both cases, the customer
service quality is reduced and time is spent trying to fix the situation.
It is generally assumed in the literature that under a decentralized decision making approach, the marketing
decisions are taken first, deciding the quantities to produce of each final item. Then, these quantities are taken
by production that decides how to produce then minimizing the total cost function. However, this is not true for
systems where both production rate and production capacity are fixed. When a new factory is set up, the
amount of capacity available is decided and this is the most critical strategic decision a producer has to make.
Surprisingly (or not), this decision concerns production, and it was made before establishing the demand.
Therefore, marketing has to deal with order acceptance (including promotion sales) respecting the limits set up
by capacity. The way to maximize revenues is to exploit the capacity usage to the maximum. To attain this
objective, the company has to use high quality optimization techniques of production planning and scheduling.
Without the possibility of changing the production capacity and maintaining a uniform production rate, the way
of dealing with unstable demand is to produce inventories anticipating future demand. An adequate frequency
of replanning is also a key issue to deal with instability of the demand (due to very flexible order management
procedures). To avoid frequent changes of plans (that give too much “noise” to the system), schedules are fixed,
for a short planning horizon, and only just before its implementation. Optimal schedules for the beginning of
the planning horizon are frequently independent of the demand data from future periods.
Mukhopadhyay and Gupta (1998) list several interface variables that can be used to resolve conflicts within an
organization. One of these variables regards information technology: online data sharing helps to reduce costs
and to maintain all the actors aware of the current data. Besides, they also suggest adopting a total quality
management philosophy, making improvements on the cost control techniques and on the shared incentives.
684
Decision Support in an Uncertain and Complex World: The IFIP TC8/WG8.3 International Conference 2004
Our experience has shown that under a decentralized structure, it is worthwhile to establish interfaces between
marketing and production through some functions. In section 4, we propose a decision-making framework
where the order acceptance problem and the forecasting of demand act as interfaces for marketing/production
coordination. It extends the decision making model we developed for the case study presented in the next
section. Although it was originally designed for a particular problem in the paper industry, it can be adapted to
other industries.
3. CASE STUDY – AN INTEGRATED DSS FOR MARKETING-PRODUCTION
DECISION-MAKING
3.1 The history of our DSS evolution
Our case study concerns an R&D experience, that lasted from more than ten years, in cooperation with the
major Portuguese paper producer (one of the world leaders in the paper market). We started to develop a
Decision Support System (DSS) for production planning and scheduling. Primarily it was only concerned to
assist production schedulers, but along the time this DSS has evolved to also assist managers. Besides all the
cost reductions achieved, the outcome is an interface between marketing and production that reduces crossfunctional conflicts.
We have worked together with two of the company mills, both producing paper rolls in a production to order
policy. Production to stock is also acceptable, namely when demand is low. Our cooperation has begun in 1989,
with the goal of designing and developing a computer tool to assist production schedulers in programming trim
decisions. As they produce several qualities of the main product, a new approach was developed that combined
the resolution of the one-dimensional cutting stock problem with orders partition and assignment of qualities.
Due to the success of the cooperation, in 1992 the software product already developed and in use was extended
to tackle the problem of capacity planning. The result was a DSS that included several features such as capacity
planning, trimming, and data selection and extraction. Near 1995, we started to cooperate with the second mill.
The proposal was to adapt and extend the existing DSS. We were asked to develop new modules to assist
marketers in the processes of order acceptance. This motivated a new study of the global problem, emphasizing
on the interdependencies of the decision-making subproblems. Although the production processes are similar
for both factories, differences on the product characteristics lead to the need of modifying the existing DSS
model base. A forecasting module was developed to assist marketing in demand analysis and medium-term
forecasting. The company was aware of the need of coordinating marketing and production functions. They
were aiming at improving customer service quality. Besides, by that time, the two units (marketing and
production) were geographically separated, operating at different cities, so increasing the difficulty of
communication. A hierarchical model with two levels of decision-making was developed. It integrated the subproblems of demand forecasting, order acceptance, capacity planning and scheduling, and cutting. Therefore,
the resulting DSS has been extended and adapted correspondingly.
Nowadays, we keep cooperating with the company. Extensions of the DSSs are envisaged. We have developed
approaches to tackle new problems, such as the bi-objective problem of sequencing cutting patterns (Respício
and Captivo 2003) and the problem of integrating short-term scheduling with capacity planning (Respício and
Captivo 2002).
3.2 Problem description
Clients order quantities of paper rolls specifying a combination of the attributes: paper type, paper basic weight,
rolls diameter and rolls width. We consider an aggregation of the final items (rolls) into families. Each pair of
the attributes paper type and paper basic weight defines a given family.
The manufacturing process is schematized in figure 1. It is organized into two stages that repeat continuously.
reels
pulp
Paper machine
(Capacity planning and
scheduling problem)
...
reels
Cutting machine
(Cutting plans
optimization problem)
...
packing
clients
warehouse
Final items
Figure 1. The manufacturing process
685
Marketing-production interface through an integrated DSS
Firstly, the paper machine produces a large stream of paper of a given family. This stream is wounded
originating a reel. Secondly, each reel is cut to obtain a given set of widths. Set-ups occur in both stages. Global
production capacity is defined on the first stage (the cutting process is faster) and therefore bottlenecks relate
with the paper machine. Final items are packaged for shipping or temporary storage. The stock consists of
items not assigned to any order.
To each manufacturing stage corresponds an optimization problem. The Capacity planning and scheduling
problem relates with the paper production and consists of assigning capacity to families’ production over a
planning horizon. The resulting plan determines the quantities to be produced of each family and the
correspondent production sequence. There are set-up times dependent on the family sequence.
The Cutting plans optimization problem relates with the process on the cutting machine and it consists on
determining how reels should be cut into items. This problem corresponds to the short-term scheduling
problem. Cutting reels of a given family and a given diameter originates items of that family and diameter. A
cutting pattern (or pattern) is a set of item widths, each of which to be cut a given number of times. A cutting
pattern is feasible if it is technically possible to cut a reel into the correspondent combination of widths. A
cutting solution is a set of feasible patterns, each of which to cut a given number of reels. The Cutting Stock
Problem consists of determining a cutting solution, producing a given set of items with minimum waste.
Assigning items in a cutting solution to client orders originates a cutting plan.
3.3 Overview of our DSS Model base
3.3.1 Assigning stock items to client orders
This function supports the computation of effective demand and precedes any process of order acceptance or
planning/scheduling. We use a priority heuristic rule following the Earliest Due Date criterion. Orders are
handled on non-decreasing order of delivery date. The procedure looks for stock items with appropriate
attributes and performs the largest possible assignment. The user is shown a report of all the assignments and
can accept it, refuse it or accept part of it, canceling the undesirable assignments.
3.3.2 Order acceptance by exchanging allocation of capacity to production orders
This function is a first step to evaluate the impact of accepting a given proposed order. The result totally
depends on the rational model of the decision maker. He/she introduces a value for the expected lost of capacity
(trim waste plus set-up waste) and the system computes the cumulative effective capacity and compares it with
aggregated cumulative demand over a discrete time horizon. The system analyses accepted orders and, if
capacity lacks, suggests a list of changes with minimal impact, including postponing delivery dates or
correcting demanded quantities. Customers are assigned a priority index and we apply rules considering
customer priorities to determinate the “most profitable” capacity allocation. This kind of procedure is widely
accepted because it allows for both objective and subjective allocation of capacity to production of orders
(Crittenden 1992). The decision maker interactively accepts or refuses changes until cumulative capacity is
sufficient to produce cumulative demand.
3.3.3 Order acceptance / capacity planning and scheduling
The problem is decomposed into two sub-problems considering different time scales. In the first one, we
determine the quantities to produce over a discrete time horizon divided into production periods, assuming each
order to be produced during the period preceding its due date. Family production quantities are determined by
obtaining a cutting solution for each family. Given a cutting solution, each pattern is disaggregated into a set of
jobs each of which cutting a minimum set of orders. The second sub-problem regards time disaggregation, and
consists of one machine batching and scheduling with family sequence dependent times (Potts and Van
Wassenhove 1992). A batch is a sequence of reels of the same family to be produced continuously, defining a
family production quantity. A job consists of cutting a set of reels, with items assigned to a given set of orders.
The optimizing criterion is the minimization of late jobs. The problem is NP-hard and we developed a heuristic
solution procedure. For each family, jobs are ordered in non-decreasing order of due date (Monma and Potts
1989). Family batches are built by grouping sub-sequences of jobs whose due date is in a given time window,
the length of which equals a half production period. The most attractive batch sequence is chosen amongst a set
of sequences generated by priority rules.
The correspondent aggregate plan displays for each production run the beginning time, the processing time, the
end time, the quantity, the cutting waste, the set-up waste and the earliest and latest due dates. It also presents
the relevant aggregated values and several economic parameters, allowing deciding on order acceptance. The
procedure also generates the correspondent disaggregated solutions.
686
Decision Support in an Uncertain and Complex World: The IFIP TC8/WG8.3 International Conference 2004
3.3.4 One-dimensional cutting stock problem
To determine a cutting plan the decision maker selects a set of orders to cut (or partial orders) of a given family
and diameter. Orders for a common width are aggregated and, the effective demand for each width is
computed. We consider an Integer Linear Programming model, based on the one of (Gilmore and Gomory
1961), including a tolerance parameter defining upper and lower bounds for the fulfillment of the demand
constraints. We solve the LP relaxation by implicit column generation. Due to the tolerance, rounding the final
LP solution by an appropriate procedure is a near optimal cutting solution.
The decision maker is offered a set of functionalities to test different scenarios, generating new solutions and
improving a given solution. Introducing new items in the plan may not increase the number of reels, thus
achieving economies of scale. This procedure is especially helpful if some patterns cut a large waste. In this
process, the decision maker uses the forecasting functionalities to determine the most adequate item widths.
3.3.5 Assignment of items to client orders
Given a cutting solution, items are assigned to client orders aiming order spread minimization. A heuristic
iteratively determines a minimum cardinality set of orders that might be cut using a given pattern. Assignment
is performed, updating the number of “not yet assigned” items. Priority is given to assignments that fulfill or
are “close to fulfill” the maximum number of orders. Short runs and out runs may occur, as a consequence of
computing non-exact optimal cutting solutions. The decision maker controls short runs while generating
cutting plans, and interactively handles out runs during the assignment process. The system suggests possible
assignments and the decision maker chooses the most appropriate.
3.3.6 Demand forecasting
The forecasting module supports marketing decisions by monitoring the evolution of paper demand, as well as
decisions related to production planning, cutting, and inventory management. Potential benefits include: lower
opportunity loss costs — which are due to the refusal of orders at times when the production maximum
capacity is attained — better usage of the plant capacity when demand is low; reducing lead times; and,
reducing cutting waste, by suggesting a larger number of alternatives to explore in the production plans.
3.4 Architecture and implementation details
The system is physically installed in personal computers assigned to different functional areas of the company,
namely the Marketing Department, the Production Department and the shop-floor (figure 2). All the computers
are connected by a data communication system, and data sharing is accomplished through the company’s
central Information System (IS). The corporate database keeps the updated information on client orders. The
DSS runs locally and the decision making procedure results are kept on the local databases. Processes called
locally perform data exchanging and allow for updates of the corporate database. The users from the Marketing
and Production Departments have permissions to access all the DSS functions. Specifying the permissions for
the different users ensures data security. Users from the Production Department do not have permissions to
access information regarding sale prices. On the other hand, users from the Marketing Department do not have
access to change the amount of planned orders. At the shop-floor, the DSS only provides access to the functions
regarding the determination of cutting plans. The purpose of using the system there is to correct errors on
pursuing the plans established by the Production Department, due to eventual failures on the cutting machine.
The successful production of orders by the cutting machine is communicated to the PC on the shop-floor and,
afterwards, to the central system, allowing for the updating of the relevant data on the corporate database.
687
Marketing-production interface through an integrated DSS
Local
Data
Base
Local
Data
Base
Corporate
Data Base
PC (Marketing Dept.)
PC (Production Dept.)
Local
Data
Base
Central system of
the company
Cutting machine
PC
(Shop-floor)
Code-bars printer
Figure 2. Part of the distributed computer architecture of the company
The users call local processes to perform the following data extracting and corresponding file import: historic
records of orders (PCLEH); records of accepted orders or proposed orders (on analysis) (PCLEP); and records
of stock items (PCSTK). These data files are updated daily and also by user request. Major changes (e.g., a
large order cancelled) that can harm the current plan are acknowledged due to the interaction between the
different decision makers (by E-mail or telephone).
Following user requests, the system generates local files for cutting plans (PCOCR), assignment of stocks to
client orders (PCAFT) and changes on active orders or proposed orders (PCCPR), as shown in figure 3. Local
processes export these files to the IS and update the databases. The cutting machine, on the shop-floor, also
receives the information exported on the cutting plan files.
IS
history
order records
order records
PCLEH
LOCAL
SYSTEM
DSS
action request
PCAFT
DM
action result
stock assignment
records
cutting plan
PCOCR
stock records
PCSTK
PCLEP
changes on orders
PCCPR
IS
Figure 3. Context diagram
The DSS supports the user interaction with the models providing several data management functionalities:
•
Information searching, extracting and ordering, according to different criteria and levels of
aggregation;
•
Order partitioning allows for building different scenarios to generate and evaluate short-term
scheduling solutions;
•
Changing order records (quantities and due dates) is helpful to perform “what-if” analysis on order
acceptance and capacity planning decision processes;
•
Monitoring and updating the DSS parameters;
•
Monitoring and updating the parameters concerning the production processes (capacity availability,
production rates, expected losses, technical constraints of the cutting machine and others).
688
Decision Support in an Uncertain and Complex World: The IFIP TC8/WG8.3 International Conference 2004
4. THE DECISION-MAKING FRAMEWORK
The decision-making framework on figure 4 describes the decision-making process, using the notation
proposed by Schneeweiss (1999).
Marketing – Order acceptance
Decision maker
Forecasting
model
Short-term
scheduling
model
Instructions (Marketing)
Production – Planning/scheduling
Decision maker
feedback(Production)
Order acceptance/
Capacity planning
model
Short-term scheduling model
Interaction
Data
Instructions (Production)
Implementation
IS
Figure 4. Decision-making framework to coordinate marketing and production
Our framework integrates the decision-making problems of order acceptance, capacity planning and short-term
scheduling. The model consists of two levels that correspond to Marketing and Production, respectively. A
single human decision maker is considered in each of these levels, although different people may take the
responsibility for the corresponding function.
Order acceptance is the decision function of the Marketing level. Pricing and promotion issues are excluded
from our approach. By doing so, the marketer attempts to fix prices as high as possible, satisfying client
requests to the largest extend. He takes the responsibility to negotiate with customers. As the capacity is fixed,
the criterion, at this level, is to maximize capacity exploitation. The Marketing decision maker interacts with
the order acceptance model (which is the same as the capacity planning and scheduling model) to evaluate the
impact of accepting a given order. In this interaction he/she is not concerned with how that order will be
produced. The objective is to determine when the production of that order will be concluded.
Capacity is limited and part of it is wasted during the production process (set-ups and wastages). Therefore, the
short-term scheduling issues influence the maximization of the capacity usage. The capacity planning and
scheduling model should integrate part of the decisions concerning the short-term scheduling.
In a situation of capacity surplus, the Marketing decision maker interacts with the forecasting model to
determine which items are more likely to be ordered in the medium-term. He/she may generate produce-tostock orders, so anticipating the production of items, and leading to a more effective exploitation of capacity.
Marketing establishes the demand to be produced giving down this information as instructions to the
production; Production reacts and responds with feedback to Marketing.
The decision maker at the Production level has to plan and schedule the shop-floor. He/she is responsible for
the operational planning. At this level, the optimizing criterion is the minimization of production costs. The
decision functions include determining capacity allocation to produce orders and establishing the short-term
schedules to be executed on the shop-floor.
To decide about capacity planning, the human scheduler interacts with the model the marketer used for order
acceptance. However, the production scheduler objective is to determine how orders should be produced. Both
decision makers share the same model – the order acceptance/capacity planning and scheduling model – to
make different kind of decisions. In that way, this model behaves as a variable interface between the two
decision makers, somehow like a mediator. The differences on the actors’ perspectives are eliminated, by
considering a common criterion: maximizing the exploitation of capacity. In fact, when capacity is fixed and
the demand is known, this criterion corresponds to minimizing production waste.
Production is also concerned with establishing short-term schedules. This function is accomplished by
interacting with the short-term scheduling model. The decision maker at Production also interacts with the
forecasting model to have insights about items that are most likely to be ordered in the short-term. Economies
689
Marketing-production interface through an integrated DSS
of scale and better exploitation of capacity may be achieved by anticipating the production for the convenient
items. Final short-term schedules are given down as instructions to be implemented by the shop-floor.
The Information System (IS) receives the outputs of the shop-floor, and performs the corresponding updates.
Both management levels access the updated information.
5. CONCLUSIONS
Our framework for decision-making under a decentralized structure relies on establishing interfaces between
marketing and production – potential conflicting functional areas. It resulted from the development of a DSS
for production planning and scheduling for a real paper producer. The system is physically installed in the
Marketing Department and in the Production Department. Data transferring is done through the computer
network of the company. The DSS integrates features to assist marketers on order management, and to assist
schedulers on elaboration of production schedules. Different decision makers, with distinct functional
perspectives, share common procedures. Use of the system has improved the coordination between the different
phases of the global decision problem. Regarding order acceptance, the large number of interaction cycles
between marketing and production was dramatically reduced. Meetings were not replaced but enhanced. The
decision makers can now concentrate on the discussion of higher-value issues. The DSS ended up promoting
cooperation and reducing conflicts. Therefore, it acts as an interface between marketing and production.
The DSS is an interactive easy to use tool, with the usual database management functions, and also with the
capacity to quickly generate high quality solutions. The user is allowed to change the solutions proposed, and to
evaluate alternatives. So, the decision maker is enabled to take advantage of his knowledge and expertise. The
system did not replace the roles of people involved. Its purpose is to help them to execute their tasks and
responsibilities with a better performance.
The benefits for the company are a global production cost reduction. On the short term, savings are due to
improving the quality of solutions, and speeding up their generation. On the long term, due to a better
organization of information regarding production, better coordination of the planning process, reductions in
lead times, and improvement of costumer service.
The company will not disclose the cost reductions or business improvements that have been achieved. The
continued usage of the DSS for over 14 years is the main indicator of its value and contribution.
REFERENCES
Amaro, G., Hendry, L. C. & Kingsman, B. (1999) Competitive advantage, customisation and a new taxonomy
for non make-to-stock companies. International Journal of Operations & Production Management, 19,
349-371.
Crittenden, V.L. (1992) Close the marketing/manufacturing gap. Sloan Management Review (pp. 41-52),
Spring.
Eliashberg, J. & Steinberg, R. (1993) Marketing-production joint decision-making in J. Eliashberg & G. L.
Lilien (Eds.), Marketing, Handbooks in Operations Research and Management Science, 5 (pp. 827-880),
North-Holland.
Gilmore P.C. & Gomory R.E. (1961) A linear programming approach to the cutting-stock problem. Operations
Research, 9, 849-859.
Keskinocak P., Wu F., Goodwin R., Murthy S., Akkiraju R., Kumaran S. & Derebail A. (1998) Scheduling
solutions for the paper industry, Research Report RC21279, IBM.
Massey G. & Dawes, P. (2000) The effectiveness of marketing relationships: an exploratory study. ANZMAC
2000: Visionary Marketing for the 21th Century: Facing the Challenge, 1492-1503, URL http://
130.195.95.71:8081/www/ANZMAC2000/ CDsite/papers/m/Massey1.pdf, Accessed 20 Nov 2003
Monma C.L. & Potts, C.N. (1989) On the complexity of scheduling with batch setup times. Operations
Research, 37 (5) , 798-804.
Mukhopadhyay, S.K. & Gupta, A.V. (1998) Interfaces for resolving marketing, manufacturing and design
conflits – a conceptual framework. European Journal of Marketing, 32, 101-124.
Murthy S., Akkiraju R., Goodwin R., Keskinocak P., Rachlin J., Wu F., Yeh J., Fuhrer R., Kumaran S.,
Agarwal A., Sturzenbecker M., Jayaraman R.& Daigle R. (1999) Cooperative multiobjective decision
support for the paper Industry. Interfaces, 29 (5), 5-30.
690
Decision Support in an Uncertain and Complex World: The IFIP TC8/WG8.3 International Conference 2004
Pickard G. & Yeager R. (1997) Paper machine scheduling yields gains at James River Naheola Mill. Pulp and
Paper, 71 (9), 125-129.
Potts C.N. & Van Wassenhove L.N. (1992) Integrating scheduling with batching and lot-sizing: a review of
algorithms and complexity. Journal of the Operational Research Society, 43 (5), 395-406.
Respício, A., Captivo, M. E. & Rodrigues, A. J. (2002) A DSS for production planning and scheduling in the
paper industry, in F. Adam, P. Brezillon, P. Humphreys, & J-C. Pomerol (Eds.), Decision making and
decision support in the Internet age (pp. 298-308), Cork, Ireland: Oak Tree Press.
Respício, A. & Captivo, M.E. (2002) Cutting stock within a production planning problem: exact solutions,
SICUP 2002 (Meeting of the Special Interest Group on Cutting and Packing) IFORS 2002, Edinburgh,
Scotland.
Respício, A. & Captivo, M. E. (2003) A bi-objective approach to sequencing cutting patterns, MIC2003
Proceedings, The Fifth Metaheuristics International Conference (pp. 62.1-62.6), Kyoto, Japan.
Schneeweiss C. (1999) Hierarchies in distributed decision making, Germany, Springer.
ACKNOWLEDGEMENTS
This research has been partially supported by the Portuguese Foundation for Science and Technology under the
POCTI program (Projects POCTI/MAT/139/2001, POCTI/MAT/2046/2001 and POCTI/152). The authors
acknowledge their industrial partner – Portucel – for the cooperation and support.
COPYRIGHT
Ana Respício & Maria Eugénia Captivo © 2004. The authors grant a non-exclusive license to publish this
document in full in the DSS2004 Conference Proceedings. This document may be published on the World
Wide Web, CD-ROM, in printed form, and on mirror sites on the World Wide Web. The authors assign to
educational institutions a non-exclusive license to use this document for personal use and in courses of
instruction provided that the article is used in full and this copyright statement is reproduced. Any other usage
is prohibited without the express permission of the authors.
691