journal modeling

8/2/2019 Journal Modeling

1/18

Theory and Methodology

A modeling approach to logistics in concurrent engineering

Shad Dowlatshahi *

Department of Information and Decision Sciences, College of Business Administration, University of Texas at El Paso, El Paso,

TX 79968-0544, USA

Received 21 July 1997; accepted 30 October 1997

Abstract

This paper advocates logistics involvement in the early phases of product design and development in a concurrent

engineering environment. A concurrent engineering environment and the benets of such involvement are explained in

detail. The paper focuses on facilitating an interface and a collaboration between the designer and the logistician. A

conceptual interface for design for logistics is presented. Four areas of interface are considered: (a) logistics engineering,

(b) manufacturing logistics, (c) design for packaging, and (d) design for transportability. A set of detailed design factors

pertaining to each area are presented. A modeling approach, Bond Energy Algorithm, is used to accomplish the design

for logistics concerns developed throughout the paper. An example is provided to test and validate the algorithm. Theresults are analyzed and appropriate perspectives for managerial implication of the methodology are provided. Finally,

some conclusions and assessments are presented. 1999 Elsevier Science B.V. All rights reserved.

Keywords: Design for logistics; Concurrent engineering; Transportability; Manufacturing product design; Bond energy

algorithm

1. Introduction

Product design in a concurrent engineering en-

vironment focuses on an interdisciplinary ap-proach that utilizes methods, procedures, and rules

to plan, analyze, select, and optimize the design of

products. In the early stages of the design process

concurrent engineering considers and includes

various product design attributes such as aes-

thetics, durability, ergonomics, interchangeability,

logistics, maintainability, marketability, man-

ufacturability, procurability, reliability, reman-

ufacturability, safety, schedulability, serviceability,

simplicity, testability, and transportability. Thegreatest impact and benets of concurrent engi-

neering are realized at the design stage of product

development. The design decisions made in the

early phases of product design and development

will have signicant impact upon future manu-

facturing and logistical activities. The following

examples illustrate the relationship.

A study at Rolls Royce revealed that design de-

termined 80% of the nal production cost of

2000 components (Corbett, 1986).

European Journal of Operational Research 115 (1999) 5976

* Tel.: 915 747 7759; fax: 915 747 5126; e-mail: sdowlats@u-

tep.edu.

0377-2217/99/$ see front matter 1999 Elsevier Science B.V. All rights reserved.

PII: S 0 3 7 7 - 2 2 1 7 ( 9 8 ) 0 0 1 8 4 - 2


2/18

According to General Motors executives, 70%

of the cost of manufacturing truck transmissions

is determined in the design stage (Whitney,

1988).

Ford Motor Company has estimated that

among the four manufacturing elements of de-

sign, material, labor, and overhead, 70% of all

production savings stem from improvements in

design (Cohodas, 1988).

A study revealed that product design is respon-

sible for only 5% of a product's cost; it can,

however, determine 75% or more of all manu-

facturing costs and 80% of a product's quality

performance (Huthwaite, 1988).

Yet another study shows that 70% of the lifecycle cost of a product is determined at the de-

sign stage. The life cycle cost here refers to cost

of materials, manufacture, use, repair, and dis-

posal of a product (Nevins and Whitney,

1989).

Fig. 1 depicts the system approach to the design

of a product. This gure provides a conceptual

framework where interactions among various

functional areas in a concurrent engineering envi-

ronment can be explored and analyzed. The

functional areas presented in Fig. 1 encompass all

relevant areas pertaining to product life cycle from

product inception to disposal. This includes busi-

ness, production, and support requirements. Each

of these requirements consists of several related

functional areas as depicted in the gure. Fig. 1

also signies that product design is at the core of

concurrent engineering activities and it provides

the most signicant benets when considered in

conjunction with any of a combination of several

functional areas. The paper focuses on the linkage

between product design and logistics as high-

lighted in Fig. 1.

1.1. Objectives and organization of the paper

The focus of this paper remains on the inte-

gration of logistics concerns in the early stages of

product design and development. The objectives of

this paper are threefold.

To demonstrate the importance of early logistics

involvement in the product design process.

To present a conceptual as well as an analyti-

cal basis for integrating logistics concerns,

Fig. 1. Product design in a concurrent engineering environment.

60 S. Dowlatshahi / European Journal of Operational Research 115 (1999) 5976


3/18

constraints, and contributions in the design pro-

cess. The conceptual and analytical bases are

not an end in themselves, but are used to facili-

tate the design for the logistics process.

To provide a framework where managerial im-

plications of design for logistics can be explored.

Furthermore, the eciency and eectiveness of

the methodology, results, and their managerial

implications are analyzed.

After a presentation of system approach to

concurrent engineering and the literature review, a

conceptual framework for design for logistics is

presented in Section 2. This section consists of a

systems approach to design for logistics. It also

addresses the components of the design for logis-tics and their associated modules and design fac-

tors. A design for logistics model is developed in

Section 3. The details of the algorithm are outlined

as well. This section will also present an example

to validate and test the approach and methodology

oered in this paper. Section 4 provides the ad-

vantages and implications as well the conclusion

and assessment.

1.2. System approach to concurrent engineering

There are two issues at the core of successful

implementation of concurrent engineering:

1. All activities related to the development of a

product should be focused in the early stages

of product design (conceptual design) so that

the greatest benets of such an integration are

achieved. The information requirements and

exchanges at the conceptual design are not

well-dened and usually fuzzy. This poses a

challenge for implementing concurrent engi-

neering.

2. The impact and constraints associated with var-

ious functional requirements should be commu-nicated to the designer on a timely, accurate,

and relevant basis.

Fig. 2 represents an integrated logistics system

as it relates to product design. An eective design

for logistics cuts across a number of functional

areas as illustrated in Fig. 2. These activities con-

verge to product design as the embodiment of all

future activities. As the design for logistics aects

Fig. 2. Integrated logistics system for product design.

S. Dowlatshahi / European Journal of Operational Research 115 (1999) 5976 61


4/18

other functional areas, other areas in turn aect

logistics considerations. This process is inherently

a dynamic one requiring negotiation and trade-o

among the functional areas in a concurrent engi-

neering environment. Dierent functional areas of

Figs. 1 and 2 and their interrelationships can be

explored and culminated into several other re-

search papers. The focus of this paper, however, is

on exploring and analyzing the link between lo-

gistics and product design.

1.3. Review of literature

A review of literature reveals that little or nowork is being done on the interface of product

design and logistics. The early and crucial interface

and collaboration between product design and

logistics has been largely disregarded. Some relat-

ed literature for the design for logistics topics are

presented in Table 1.

The review of literature indicates that no direct,

signicant work has been done in bridging the gap

between product design and logistics. It is impor-

tant to note that most papers focused on the

qualitative and non-numerical evaluation of lo-gistics. There was only one paper that focused on

the quantitative and analytical evaluation of lo-

gistics manufacturing within a global manufac-

turing environment. This paper, however, did not

address the crucial logistics factors in the early

stages of product design. Additionally, it only

considered the role of logistics manufacturing and

not other factors such as logistics engineering,

packaging, and transportability issues. This paper

combines conceptual and quantitative approaches

to design for logistics by rst developing a con-

ceptual framework and then by using a modelingapproach to create self-contained clusters that are

easily manageable and can be implemented si-

multaneously.

2. A conceptual framework to design for logistics

A system approach to design for logistics es-

sentially includes a designer's functional require-

ments as well as a logistician's requirements of

availability, supportability, cost, quality, volume

changes, timely delivery, order frequencies, and

the like. The design for logistics in Fig. 3 is de-

composed into four subsystems.

Fig. 3 suggests that the design for logistics has

four essential subsystems whose presence and in-

teractions largely determines the content of design

for logistics. These subsystems encompass design,

manufacturing, marketing, logistics, distribution,

packaging and other similar aspects of the overall

product design. The scope of this paper does not

include design for environment and recycling is-

sues. In other words, these four subsystems are the

embodiment of the functional areas presented in

Fig. 2.To achieve exibility, design economics, and

overall design optimization, the design for logistics

presented in Fig. 3 needs to be decomposed fur-

ther into manageable and homogeneous units. For

simplicity purposes, these manageable and homo-

geneous units are called modules. Therefore, each

subsystem of Fig. 3 is, in turn, divided into mod-

ules. Each subsystem can be composed of several

modules. Modules are the building blocks of de-

sign for logistics. Each module is further decom-

posed into design factors. Each module, then, hasseveral respective design factors that must be

considered in the design of that particular module.

Fig. 4 shows the entire decomposition process

of design for logistics and, as such, provides a

conceptual framework for the inclusion of logistics

concerns, constraints, and contributions in the

design process. Each module addresses a specic,

manageable and homogenous aspect of design for

logistics. The logic is that modules are easier to

develop, manage, and implement than a mono-

lithic system with all the complexity it oers. The

modules presented in Fig. 4 are intended to begeneric, yet comprehensive, for the design for lo-

gistics. The composition, nature, and the number

of modules may be modied (new ones may be

added or the existing ones be deleted) based on the

unique requirements and the complexity of each

design for logistics environment. One may argu-

ably question the necessity of having all the

modules presented in Fig. 4 or judiciously propose

adding new ones. These decisions are largely de-

pendent upon the discretion of the designer and



5/18

based on the managerial implications of each

manufacturing organization. Notwithstanding this

statement, Fig. 4 may be used as a suitable con-

ceptual framework for any design for logistics of

mechanical products. Additionally, design factors

are the driving force behind the modules. Design

factors are the smallest functional requirements in

the overall design of logistics. The design factors in

Table 1

Review of literature

Reference Focus/theme Methodology

Foo et al. (1990) Product design from a materials logistics standpoint. The ideal product has:

the minimum number of parts, standard or ``preferred'' parts, a modular

and reusable physical architecture, a limited set of end item congurations,

and a modular BOM structure.

Conceptual

Starsmore (1990) The eect of design decisions on logistics operations of corrosion

inhibition. The design decisions spanned from conception

to decommissioning of the product.

Conceptual

Colling (1991) The role of material selection in manufacturing processes, quality control,

product packaging, product failure and product safety was discussed.

Conceptual

Kyng (1991) Informational issues related to an integrated logistics system.

Interoperability and collaboration with regard to material accessibility.

Conceptual

Implementing Concurrent

Engineering for SMT (1992)

Surface Mount Technology (SMT), in order to be successful, requires

the implementation of both concurrent engineering and design for

manufacturability.

Conceptual

Mather (1992) The basis for design for logistics (DFL) is the rm's ability to give a

customer the required products when wanted. It is imperative that the

manufacturing processes change (as needed) to suit an environment in

which the manufacturing changes can be most eective.

Conceptual

Fawcett and Closs (1993) US manufacturers must do more than improve their internal eciency

and eectiveness. They must also develop an understanding of how to

better structure productive resources. A discussion of the ndings of a

path analytic investigation of the interrelationships between a rm's

perception of economic globalization, a rm's emphasis of logistics and

manufacturing considerations in the design and management of global

manufacturing networks, and a rm's competitive/nancial performance.

Conceptual and

quantitative (path

analysis)

Petersen (1993) Industrial engineers are the ideal professionals to design and integrate

logistics, transportation and distribution systems.

Conceptual

Germain et al. (1994) The relationships between logistics technology adoption and

organizational design practices are examined and the logistics

technologies of several organizations are discussed.

Conceptual and

empirical (36 logis-

tics technologies)

Cherf (1995) For factories with design capabilities, a plan to reduce cost with design

for manufacturability and documentation is proposed.

Conceptual

``A Modeling Approach to

Logistics in Concurrent

Engineering'' (1997)

A comprehensive hierarchical structure of design for logistics is

presented. An algorithm is used to model logistics problems in

concurrent engineering.

Conceptual and

quantitative (Bond

Energy algorithm)



6/18

Fig. 3. Components of design for logistics.

Fig. 4. Hierarchical decomposition of design for logistics.



7/18

a logistics design are synonymous to ``parts'' in

product design.

Each subsystem of the design for logistics is

subsequently explained and design factors associ-

ated with each module are delineated.

2.1. Logistics engineering

Logistics engineering is a eld of logistics that

deals with the supportability of product and sys-

tems throughout their life cycle. Logistics engi-

neering is concerned with the design process in

that it establishes requirements to which the ulti-

mate design conguration must comply. Logisticsmust be intuitively an integral part of the design

process along with performance, size and weight,

reliability, safety, manufacturability, cost, and the

like. Logistics engineering also takes into consid-

eration environmental impact, energy conserva-

tion, and solid waste transportation and disposal.

Design and logistics engineers should collaborate

by having sound and complete design specica-

tions, periodic visits, eective and continuing

communication and dialogue, and development of

eective planning documents.This area addresses system, product design, and

development. Logistics engineering criteria sup-

port the life-cycle aspects of product design (design

for product supportability). The focus remains on

inuencing and incorporating changes in the

product design to meet product supportability

requirements before the design is nalized. The

design factors associated with the logistics engi-

neering subsystem are presented in Table 2.

The list of design factors presented in Table 2

(and subsequent Tables 35) are not intended to

be exhaustive. These generic lists merely presentthe design factors that may be appropriate to use

in the design of these subsystems. Additional

design factors may be added or deleted from

Tables 25 depending on the unique requirements

and complexity of each logistics systems design.

Notwithstanding this statement, Tables 25 still

present the most comprehensive list of design

factors developed for the modules specied in

Fig. 4. Also, noteworthy is that there are no pri-

orities in the ordering of these design factors, nor

is any design factor more important than any

other. Only the designer's experience and judg-

ment, in terms of a factor's relevance or impor-

tance, determines the inclusion and/or exclusion of

certain design factors in Tables 25.

2.2. Manufacturing logistics

The characteristics of the manufacturing pro-

cesses and activities are a major determinant of the

logistics activities and logistics system design.

Manufacturing processes and activities often cre-

ate a number of constraints and opportunities for

a logistics system. Manufacturability, as a main

feature of product design, is a signicant con-

tributor to the design for logistics. The design

factors associated with the modules of the manu-

facturing engineering subsystem are presented in

Table 3.

2.3. Design for packaging

This area addresses the issues related to pack-aging requirements in the product design process.

Packaging is an important feature of a product as

it creates or enhances the product's image. Product

packaging is an important marketing tool and has

a signicant impact upon overall product cost, its

ease of use, and its perception to customers.

Packaging also protects the product from break-

age, spillage, etc. Logistics requirements for

packaging must be incorporated at the design

stage with those of marketing and manufacturing

requirements. In the past, designers have at-

tempted to incorporate only marketing and man-ufacturing requirements in the design instead of

including the logistics requirements as well. This

has resulted in an ineciency in the overall prod-

uct and system performance, as well as higher

operational and usage costs. Poor packaging re-

sults in lower sales, damaged contents, customer

dissatisfaction, and higher cost of material han-

dling, warehousing and transportation. Packaging

absorbs approximately 12% of the logistics dollar

(Ballou, 1987). The design factors associated with



8/18

the modules of the design for packaging subsystem

are presented in Table 4.

2.4. Design for transportability

Transportation costs represent the most single

important element in logistics costs for most rms.

An eective design for transportability stimulates

direct competition among rms at dierent loca-

tions; creates greater economies of scale; and re-

duces the price of goods and services. The design

factors associated with the modules of the design for

transportability subsystem are presented in Table 5.

3. A modeling approach to design for logistics

This paper utilizes Bond Energy Algorithm

(BEA) a clustering approach to the design of

Table 2

Hierarchical structure of logistics engineering

Module Design factors Notation

Design for supportability (a1) Future redesigns (a1X1)

Product variation (a1X2)

Logistics performance data (a1X3)

Product performance (a1X4)

Product weight (a1X5)

Design specications (a1X6)

Operating height (a1X7)

Market characteristics (a1X8)

Storage and movement systems (a1X9)

Equipment type (a1X10)

Carrier type (a1X11)

Product test and evaluation (a1X12)

Self maintenance (a1X13)

Design for manufacturability (a2) Standardized parts (a2X1)

Part counts (a2X2)

Interchangeable parts (a2X3)

Product physical dimensions and bulk (a2X4)

Product value (a2X5)

Kinematics (type, direction of motion, velocity, acceleration, etc.) (a2X6)

Forces (direction, magnitude, and frequency of force) (a2X7)

Ease of assembly (a2X8)

Type of manufacturing processes (a2X9)

Tolerances (a2X10)

Surface nish (a2X11)

Product lines (a3) Number of changes in product lines (a3X1)

Number of product lines (a3X2)

Conguration management (a3X3)

Demand variability (a3X4)

Design leadtimes (a3X5)

Demand forecast error (a3X6)

Product seasonality (a3X7)

Product prot margin (a3X8)

Design attributes (a4) Duration of downtime (a4X1)

Measures of availability (a4X2)

Mean time to rst failure (a4X3)

Mean idle time (a4X4)

Failure rate per unit time (a4X5)

Service time (a4X6)

Time to respond to service (a4X7)



9/18

Table 3

Hierarchical structure of manufacturing logistics


Manufacturing processes (a5) Robustness of product design (a5X1)

Manufacturing tooling (a5X2)

Critical process (a5X3)

Size and dimensions of WIP (a5X4)

Utilization rate (a5X5)

Degree of automation (a5X6)

Volume exibility (a5X7)

Product exibility (a5X8)

Stressstrain analysis (a5X9)

Shape, size, and thickness of the parts (a5X10)

Tolerances and surface nish requirements (a5X11)

Production volume (a5X12)

Product expected service life (a5X13)

Type of manufacturing process used (a5X14) Assembly operations (a5X15)

Production planning and control (a6) Length of production run (a6X1)

Size of production run (a6X2)

Leadtimes (a6X3)

Production schedules (a6X4)

Set-up times (a6X5)

Throughput time (a6X6)

Inventory turnover rate (a6X7)

Seasonal inventory (a6X8)

Idle times (machine and labor) (a6X9)

Due dates (earliest, latest) (a6X10)

Slack times (a6X11)

Inventory level (a6X12) Loading (a6X13)

Average ow time (a6X14)

Average number of jobs in the system (a6X15)

Changeover cost (a6X16)

Inventory levels (a6X17)

Work in process (a6X18)

Materials (a7) Material delivery time (a7X1)

Critical and proprietary materials (a7X2)

Unit load size (a7X3)

Part type (unit, bulk, liquid, gas) (a7X4)

Piece rate (quantity) (a7X5)

Part orientation (design) (a7X6)

Logistics of the materials move (internal, external) (a7X7)

Material properties (a7X8)

Material availability (a7X9)

Material cost (a7X10)

Plant location (a8) Plant capacity (a8X1)

Demand schedules (a8X2)

Number of plants (a8X3)

Location of plants (a8X4)

Distribution of costs (a8X5)

Multiple warehouses (a8X6)

Transportation costs (a8X7)



10/18

logistics systems. The clustering algorithms have

been widely used in the decomposition of complex

design problems. The purpose of clustering algo-

rithms is to decompose a large system into sub-

systems and then cluster them into manageable

modules and activities. BEA is concerned with the

grouping of objects into homogeneous clusters

(groups) based on common features or activities.

Clustering Analysis Methods have been widely

used in various disciplines such as biology, data

recognition, medicine, pattern recognition, pro-

duction ow analysis, task selection, control

engineering, automated systems, market segmen-

tation, and expert systems.

Table 4

Hierarchical structure of design for packaging


Packaging materials (a9) Strength of material (a9X1)

Weight per cubic measure (a9X2)

Packaging density (a9X3)

Transportation rate (a9X4)

Cost of materials (a9X5)

Cost of processing the package's materials (a9X6)

Type of materials used (a9X7)

Amount of materials reduced (a9X8)

Material disposability and reusability (a9X9)

Packaging testing (a10) Shock levels (a10X1)

Moisture (a10X2)

Heat resistance (a10X3)

Vibrations (a10X4)

Corrosiveness (a10X5)

Pressure (a10X6)

Tension (a10X7)

Compression strength (a10X8)

Impact (a10X9)

Allowable level of damage/protection (a10X10)

Material fragility (a10X11)

Packaging design features (a11) Packaging shape, size, and modules (a11X1)

Packaging overall cost (a11X2)

Loss and damage historical data (a11X3)

Packaging specications (a11X4)

Packaging interior and exterior factors (a11X5)

Packaging ease of opening, closing, and reusing (a11X6)

Packaging ease of handling (a11X7)

Package identication (a11X8)

Amount of dead space in stacking (a11X9)

Weight-to-protection shipping ratio (a11X10)

Temper-resistant packaging (a11X11)

Functional packaging requirements (a12) Packaging physical dimensions (a12X1)

Element environmental factors (temperature, humidity, etc.) (a12X2)

Cubic utilization rate (a12X3)

Packaging shape and structure (a12X4)

Inventory requirements (a12X5)

Transportation requirements (a12X6)

Warehousing requirements (a12X7)

Shipping and handling requirements (a12X8)

Order picking requirement (a12X9)

Linear footage of shelf space (a12X10)



11/18

3.1. Solution procedure for Bond Energy Algorithm

BEA is an interchange clustering algorithm that

seeks to create a block diagonal form by maxi-

mizing some measure of eectiveness. The purpose

of BEA, in general, is to identify and display

natural variable clusters that occur in complex

data arrays. In particular, the objective of BEA in

a logistics design is to group design factors into

Design Factor Families (DFF) and modules into

Table 5

Hierarchical structure of design for transportability


Transportation mode (a13) Transit time (speed) (a13X1)

Timely deliveries (dependability) (a13X2)

Number of carriers (a13X3)

The length of transportation contracts (a13X4)

Transportation mode of operational availability (a13X5)

Types of goods (sensitive, fragile, perishable, gas, liquid, etc.) (a13X6)

Size of shipment (a13X7)

Length of shipment (a13X8)

Degree of vibration and acceleration (a13X9)

Sectionalization and disassembly capacity (a13X10)

Frequency of use (a13X11)

Transportation (freight) rate for weight/mode/class (a13X12)

Wage rate (a13X13)

Exchange rate (a13X14) Trac rate (a13X15)

Product density (a13X16)

Transportation economy (a13X17)

Transportation versatility (a13X18)

Capacity (a13X19)

Design criteria (a14) Average transit times (a14X1)

Delivery time variability (a14X2)

Product time to market (a14X3)

Product value to weight ratios (a14X4)

Transit-time variability (a14X5)

Average delivery times (a14X6)

Degree of protection provided (a14X7)

Low storage cost (a14X8) High movement cost (a14X9)

Low-value weight ratio (a14X10)

Total sales (a14X11)

Transportability issues (a15) Transport method (a15X1)

Economic storing, handling, and shipping (a15X2)

Safe storing, handling, and shipping (a15X3)

Work-in-process handling (a15X4)

Physical properties (width, height, length, center of gravity, etc.) (a15X5)

Dynamic limitations (acceleration, vibration, deection, leaking etc.) (a15X6)

Life cycle cost (a15X7)

Maximum weights (a15X8)

Environmental limitations (temperature, pressure, humidity etc.) (a15X9)

Hazardous eects (radiation, explosives, electrostatic, personal safety, etc.) (a15X10)

Product value (a15X11)

Type of packaging (a15X12)

Product weight (a15X13)

Shipping and handling characteristics (weight, bulk, compactness, etc) (a15X14)



12/18

Module Families (MF). The forming of these

families allows a designer to simultaneously con-

sider and design the design factors common to a

set of modules. This modularized approach in-

creases the eciency of the logistics design.

The interactions between modules and design

factors can be represented in a binary module-de-

sign factor incidence matrix. In the formulation of

the incidence matrix [aij], the entries of 0 and 1 are

used. Entry 1 (0) indicates that a particular design

factor does (does not) belong to a module. The

binary assignment is accomplished by ``the de-

signer''. The designer may be interpreted as one

expert or a team of experts. No attempt is made to

dene the composition or the working dynamics ofthe teams. This topic is beyond the scope of this

paper. The topic of team decision-making and

group negotiations has received considerable at-

tention in current literature. The much talked

about rating and assignment techniques (such as

the Delphi method) are applicable to this paper as

well. The assignment of binary values is inherently

a technical task and requires the possession of

knowledge and skills as they pertain to a particular

logistics design. The methodology presented here

allows the generation of multiple incidence matri-ces. Each of these matrices can be explored and

solved, and the results can be evaluated for their

eectiveness. On the other hand, discussions can

be held by the design team members and changes

can be made to only one incidence matrix until a

consensus is achieved. The opportunities for de-

veloping sound incidence matrix(ces) abound.

The Bond Energy Algorithm shown in Fig. 5 is

an enhancement of the work by McCormick et al.

(1972). Fig. 5 outlines the basic BEA algorithm by

presenting two basic procedures for the columns

and rows. The eciency of this algorithm is en-hanced since the procedure used for the columns

and rows is virtually similar.

The measure of eectiveness for BEA is calcu-

lated as follows:

ME 1

2

m

i1

n

j1

ij iYj1 iYj1 i1Yj i1Yj

X

The ME measures the density of the bond be-

tween any two elements of the matrix. The greater

the numerical value of ME, the greater the bond

between the elements. The purpose of this algo-

rithm is to permutate the rows and columns and

place elements adjacent to one another which have

the largest ME values.

By using the BEA, an unstructured incidence

matrix can be transformed into a new structured

Fig. 5. Flow chart of BE algorithm.



13/18

matrix with a clear grouping of activities. The

clusters (mutually exclusive submatrices) obtained

are considered the building blocks of the logistics

system design.

For procedural simplicity only two design fac-

tors per module were selected for this paper. The

incidence matrix (1)(Fig. 6) represents the assign-

ment of design factors to modules. In the incidence

matrix (1), the aij entries correspond to design

factors presented in Tables 25. Entry 1 signies

that the inclusion of a particular design factor in a

module is a necessary and essential requirement of

forming that module. Each module inherently

consists of a set of cohesive and bounded design

factors whose interactions determine the overalldesign and eectiveness of the module. The selec-

tion and assignment of design factors to the ap-

propriate modules pose a challenge as well as an

opportunity for the designer. These decisions are

virtually interrelated and must be made with cau-

tion. In assigning the binary values of 0 and 1, two

issues must be considered. First, not all binary

decisions require a similar level of diculty, skills,

and judgment. Some binary assignments are fairly

obvious and easy to make. Secondly, the specic

requirements of each logistics design, the com-plexity of the design process, and nally the

number of design factors and modules involved

largely determine the nature of the binary deci-

sions.

Justications can be made for the binary as-

signments in the incidence matrix (1). A few ex-

amples illustrate this situation. The design factor

a13X1 (transit time) is assigned to a module 13

(transportation mode) and module 15 (transport-

ability requirements). These assignments are fairly

obvious and easy to make. By the same logic, the

design factor a1X5 (product weight) is assigned tomodule 1 (design for supportability), module 12

(functional packaging requirements), and module

13 (transportation mode) because it is a necessary

and essential design factor for forming these

modules. On the other hand, for example, assign-

ment of the design factor a9X7 (human power: time,

skill) to module 2 (design for manufacturability)

requires justication and assumption, as it is not

easily apparent as to why a9X7 is a necessary and

essential requirement of forming module 2. The

designer may conclude that the type of materials

used is directly related to design for manufactur-

ability concerns. The materials selected in the de-

sign stage need to be fabricated or assembled as a

part of the overall manufacturing processes. Cer-

tain materials (e.g. thin, fragile, and brittle com-

ponents and materials) may require special

equipment, tools, and processes. Although these

considerations are important in many operations,

they may not be necessary and essential for as-

signment of a design factor to a module. Designers

ought to make a decision on these cases by using

their skills and judgment. By so doing, the crucial

decisions are deliberately made at the design stage

and not by default at the implementation stage.Furthermore, the design factors outlined in Ta-

bles 25 may vary from one design to another

depending upon the unique requirements of each

logistics system design. Care, however, must be

taken that only the most appropriate and relevant

design factors are employed in any particular de-

sign.

The incidence matrix (1) is an unstructured

matrix that will be transformed into a new struc-

tured matrix by the use of the BEA. First, column

1 of incidence matrix (1) is selected arbitrarily andthe measure of eectiveness for each column is

calculated and presented in Table 6. From Ta-

ble 6, columns 21 and 26 with the ME value of 5

(the highest ME value) are placed next to column

1. This procedure is repeated until all columns are

accounted for and placed with one another. The

results are represented in incidence matrix (2)

Fig. 7.

By the use of the BE algorithm and selecting

row 1 arbitrarily, the measure of eectiveness for

all other rows are calculated and presented in

Table 7.Rows with the largest ME values are placed

together. From Table 7, rows 13 and 15 with the

ME value of 8 and 6, respectively are placed next

to row 1. This procedure is repeated until all rows

are accounted for and placed with one another.

The detailed solution for columns and rows are

available from the author. The combined results of

both columns and rows for the BEA are present-

ed in incidence matrix (3) is shown in Fig. 8.

The clusters (mutually exclusive submatrices) in



14/18

Fig.

7.Incidencematrix(2).

Fig.

6.

Incidencematrix(1).



15/18

incidence matrix (3) present a structured matrix

with a clear grouping of activities.

The incidence matrix (3) indicates that the

overall design of logistics can be accomplished in

four self-contained modules. Module Family 1

addresses transportation issues. It consists of

design for supportability, functional packaging

requirements, the transportation mode, and

transportability issues. Module Family 2 addresses

manufacturing issues. It consists of design for

manufacturability, materials, and manufacturing

processes. Furthermore, Module Family 3 and

Module Family 4 address production planning and

control and design characteristics, respectively.

For example, Module Family 2 focuses on theissues and problems related to manufacturability,

materials, and the actual manufacturing processes.

In designing this module family, the designer

considers only the necessary and relevant design

factors. By so doing, the scope and degree of

complexity of the design become more manage-

able. The logistics designer simply focuses on such

design factors as: standardized parts, type of

manufacturing processes, production volume,

material properties, type of materials used, and

physical properties (width, height, length, center ofgravity, etc.). As is evident, this module family not

only considers the commonly used manufacturing

design factors, but also cuts across other modules

(manufacturing processes a5, packaging materi-

als a9, transportability issues a15) and includes

other relevant and necessary design factors. This

simply signies that the manufacturability module

is not an isolated activity and may not be designed

in a vacuum-like environment. On the contrary,

this modular approach enhances the eectiveness

of the manufacturability module by reaching

and incorporating other relevant and essentialdesign factors which may have otherwise been

disregarded. The design eectiveness is further

enhanced by the manageability and ease of im-

plementation of such a modular design.

Similarly, one may consider Module Family

3 which combines ve separate modules from

three dierent subsystems of the logistics design.

This module family proposes that logistics engi-

neering, manufacturing logistics, and design for

packaging are interrelated and must be considered

Table 7

ME values for row 1

Position ME

j1 j + 1 Value

Row number 1 2 0

1 3 01 4 0

1 5 0

1 6 0

1 7 0

1 8 0

1 9 0

1 10 0

1 11 0

1 12 4

1 13 8

1 14 0

1 15 6

Table 6

ME values for column 1

Position ME

j1 j + 1 Value

Column number 1 2 4

1 3 0

1 4 0

1 5 0

1 6 0

1 7 3

1 8 2

1 9 0

1 10 0

1 11 0

1 11 0

1 12 0

1 13 31 14 0

1 15 0

1 16 0

1 17 4

1 18 0

1 19 0

1 20 0

1 21 5

1 22 0

1 23 2

1 24 2

1 25 2

1 26 51 27 0

1 28 0

1 29 3

1 30 0



16/18

simultaneously. By the same logic, Module Fami-

lies 1 and 4 may be analyzed and explained.

4. Advantages and implications

The following represents a summary of advan-

tages and practical managerial implications of the

methodology of design for logistics.

1. The methodology presented here allows the

designer to be an active participant in the design of

logistics systems. The initial incidence matrix is

developed by the designer. The nal solution to

the problem is dependent upon the initial matrix.

The designer is an inseparable part of the identi-cation and development of subsystems, modules,

and design factors and, for that matter, the entire

logistics decision-making process. The designer

has the exibility to select and assign design fac-

tors to modules based upon a designer's judgment,

skills, and the unique requirements of each logis-

tics design. In this methodology, the designer

controls the parameters and the direction of the

problem at hand and not a predetermined and

rigid model.

2. Various solutions can be developed based ondiering initial solutions. This capability enhances

the eectiveness, as well as the eciency of the

solution procedure, and provides a great deal of

exibility for the designer.

3. This methodology is applicable to matrices of

any size or shape. The only requirement is that the

elements of a matrix be non-negative. The BEA

solutions are nite. This makes the BEA algorithm

applicable to new designs as well as currently ex-

isting designs. The nal solution obtained by using

this algorithm is independent of the order in which

the rows and columns are presented.4. This methodology generates modules that are

cohesive, bounded, or contain a self-contained

group of activities. For eective implementation of

integrated logistics design, each module solves one

clearly dened segment of the total system. Mod-

ularity, as a basic rule of good design, is more

easily changed, expanded, or contracted than

monolithic system designs. Most managers nd

modular system designs easier to understand and

apply. These module families can be designed and

Fig.

8.I

ncidencematrix(3).



17/18

implemented simultaneously. This can potentially

reduce the total design cycle time and bring about

the advantages of concurrent engineering to the

design of integrated logistics.

5. The analytical approach presented in this

paper is capable of managing and organizing a

large number of design factors involved in the

design of integrated logistics with considerable

ease and less computer CPU time. The computa-

tion time, which depends solely on the size of the

matrix, is usually minimal and insignicant. The

simple program written to run this example is

available upon request from the author. The BEA

is fairly easy to apply and understand by logis-

ticians and higher level management.6. The design of large-scale manufacturing

systems such an logistics design usually requires

numerous components that must be expressed by

specications. There may be a variety of speci-

cations necessary to provide the guidance and

controls associated with the development of the

system and its components (Blanchard, 1991).

The development of these specications and cri-

teria can be eectively accomplished by modular

design. The undertaking of a very large-scale

system as a whole, with all its complex compo-nents, may be inecient or, at the very least,

cumbersome.

7. The BEA allows the inclusion of values other

than 0 and 1 in the incidence matrix (1). The de-

sign factors in this paper are assumed to be re-

duced to their lowest format allowing for no

further decomposition. It is, however, possible to

consider higher importance to certain combina-

tions of module-design factors by assigning a

higher value than 1 to these combinations. The

BEA is equipped to handle these situations well

since the placement of the adjacent columns androws are based on the ME values, and since the

ME values use aij in their calculations.

The preceding points provide some advantages

and managerial implications for the use of the

methodology of design for logistics. The selection

of any design tool, however, is largely dependent

upon the objectives and complexities of the logis-

tics design system. This selection is also aected by

the level of expertise and willingness of the de-

signer to actively participate in the process.

4.1. Conclusion and assessment

A concurrent engineering environment provides

a suitable venue to consider logistics problems

since logistics concerns, contributions, and con-

straints are best addressed in the early phases of

the product design cycle. The advantages outlined

for concurrent engineering earlier in this paper can

be realized if logistics requirements, as a part of

the overall product design, are considered.

The conceptual framework provides for an ef-

fective tool for the logistician to include the nec-

essary and relevant subsystems, modules, and

design factors. This approach allows the designer

to become a full participant in the logistics systemsdesign.

BEA provides an ecient clustering algorithm

for solving logistics problems. This algorithm

generates self-contained clusters that can be more

easily understood, changed, and implemented.

This feature allows for a logistics system to be

implemented more eectively and in a shorter pe-

riod of time since the independent clusters need

not be implemented sequentially. Fig. 9 illustrates

how the four module-families obtained in inci-

dence matrix (3) can be overlapped to reduce thedesign cycle time. Note that the bars are not drawn

to scale.

This paper has specically explored a large

number of areas where collaboration and interface

of logistics and design activities can result in

signicant achievements for a manufacturing

Fig. 9. The Gantt chart of overlapping module families.



18/18

enterprise. Although these areas of collaboration

are equally applicable to a variety of manufac-

turing logistics, care must be taken that a specic,

relevant, and custom made program resulting

from the unique requirements and environment of

each rm is selected. This process allows for a

better focus on the ensuing issues that require

more immediate attention. The most important

and essential prerequisite for an interface between

logistics and design, however, remains to be the

elimination of the ``over-the-wall-design'' concept.

All collaboration and interface must be done on a

long-term basis unless circumstances dictate oth-

erwise. Logisticians should be given an essential

role as the key player in the design process. Thisprocess is certain to fail, if legitimate authority and

power is not delegated to the logistics function.

Top management should genuinely encourage lo-

gistics involvement. If the design process is viewed

as an abstract and vacuum-like process that is

performed sequentially rather than concurrently,

no signicant achievements are expected to occur.

The eective dialogue between logistics and design

can only occur when the barriers and walls

whether real or imaginary are removed. The

support of top management and a positive insti-tutional culture are essential to instill and foster

such an environment.

References

Ballou, R.H., 1987. Basic Business Logistics, 2nd edition.

Prentice-Hall, Englewood Clis, NJ.

Blanchard, B.S., 1991. System Engineering Management.

Wiley, New York, p. 65.

Cherf, R.M., 1995. Reducing board costs. Circuits Assembly 6

(3), 82.

Cohodas, M.J., 1988. Make the most of supplier know-how.

Electronics Purchasing, July, pp. 3839.Colling, D.A., 1991. Materials and product safety. Professional

Safety 34 (4), 125.

Corbett, J., 1986. Design for economic manufacture. Annals of

C.I.R.P. 35 (1), 93.

Implementing Concurrent Engineering for SMT, 1992. Elec-

tronic Packaging and Production, August, p. 22.

Fawcett, S.E., Closs, D.J., 1993. Coordinate global manufac-

turing: The logistics/manufacturing interaction, and rm

performance. Journal of Business Logistics 14 (1).

Foo, G., Clancy, J.P., Kinney, L.E., Lindemulder, C.R., 1990.

Design for material logistics. AT&T Technical Journal 69

(3), 61.

Germain, R., Droge, C., Daugherty, P.J., 1994. A cost and

impact typology of logistics technology and the eect oftheir adoption on organizational practice. Journal of

Business Logistics 15 (2), 227.

Huthwaite, B., 1988. Designing in quality. Quality 27 (11), 34

35.

Kyng, M., 1991. Designing for cooperation: Cooperating in

design. Communications of the ACM 34 (12), 64.

Mather, H., 1992. Design for logistics (DFL) The next

challenge for designers. Production and Inventory Manage-

ment Journal 33 (1), 710.

McCormick, W.T., Jr., Schweitzer, P.J., White, T.W., 1972.

Problem decomposition and data reorganization by a

clustering technique. Operations Research 20 (5), 993

1009.

Nevins, J.L., Whitney, D.E., 1989. Concurrent Design of

Products and Processes. McGraw-Hill, New York, pp.

23.

Petersen, H.C., 1993. Logistics, transportation and distribution:

Engineering service and prots. Industrial Engineering 25

(12), 21.

Starsmore, R.P., 1990. History of a wet gas transportation

pipeline from design through decommissioning. Corrosion

Prevention and Control 37 (5), 117.

Whitney, D.E., 1988. Manufacture by design. Harvard Business

Review, JulyAugust, pp. 8391.


journal modeling

Documents