Big Data Analytics Using Multiple Criteria

Decision-Making Models 1st Edition Ramakrishnan

Ramanathan pdf download
https://textbookfull.com/product/big-data-analytics-using-multiple-criteria-decision-making-
models-1st-edition-ramakrishnan-ramanathan/

★★★★★ 4.7/5.0 (27 reviews) ✓ 228 downloads ■ TOP RATED

"Great resource, downloaded instantly. Thank you!" - Lisa K.

DOWNLOAD EBOOK
 Big Data Analytics Using Multiple Criteria Decision-Making
Models 1st Edition Ramakrishnan Ramanathan pdf download

TEXTBOOK EBOOK TEXTBOOK FULL

Available Formats

■ PDF eBook Study Guide TextBook

EXCLUSIVE 2025 EDUCATIONAL COLLECTION - LIMITED TIME

INSTANT DOWNLOAD VIEW LIBRARY

Collection Highlights

Mathematical Models For Decision Making With Multiple

Perspectives 1st Edition Gomes

Investment Decision making Using Optional Models 1st

Edition

New Concepts and Trends of Hybrid Multiple Criteria

Decision Making 1st Edition Gwo-Hshiung Tzeng

New Perspectives in Multiple Criteria Decision Making

Innovative Applications and Case Studies Michalis Doumpos
Multiple Criteria Decision Making Applications in
Environmentally Conscious Manufacturing and Product
Recovery 1st Edition Surendra M. Gupta

Business Analytics Data Analysis Decision Making 6th

Edition S. Christian Albright

Digital Analytics Data Driven Decision Making in Digital

World 1st Edition Jumin Kamki

Business Analytics: Data Analysis & Decision Making

(MindTap Course List) 7th Edition Albright

Multiple Criteria Decision Analysis State of the Art

Surveys 2nd Edition Salvatore Greco
Big Data Analytics Using Multiple
Criteria Decision-Making Models
EDITED BY
Big Data Ramakrishnan Ramanathan
Muthu Mathirajan
A. Ravi Ravindran
MCDM Tools

Business Analytics
Tools

Analytics Reports Human Interpretation

Business Intelligence
Big Data Analytics
Using Multiple
Criteria Decision-
Making Models
 The Operations Research Series
Series Editor: A. Ravi Ravindran
Professor, Department of Industrial and Manufacturing Engineering
The Pennsylvania State University, University Park, PA

Published Titles
Multiple Criteria Decision Analysis for Industrial Engineering:
Methodology and Applications
Gerald William Evans
Multiple Criteria Decision Making in Supply Chain Management
A. Ravi Ravindran
Operations Planning: Mixed Integer Optimization Models
Joseph Geunes
Introduction to Linear Optimization and Extensions with MATLAB ®
Roy H. Kwon
Supply Chain Engineering: Models and Applications
A. Ravi Ravindran & Donald Paul Warsing
Analysis of Queues: Methods and Applications
Natarajan Gautam
Integer Programming: Theory and Practice
John K. Karlof
Operations Research and Management Science Handbook
A. Ravi Ravindran
Operations Research Applications
A. Ravi Ravindran
Operations Research: A Practical Introduction
Michael W. Carter & Camille C. Price
Operations Research Calculations Handbook, Second Edition
Dennis Blumenfeld
Operations Research Methodologies
A. Ravi Ravindran
Probability Models in Operations Research
C. Richard Cassady & Joel A. Nachlas
 Big Data Analytics
Using Multiple
Criteria Decision-
Making Models

Edited by
Ramakrishnan Ramanathan,
Muthu Mathirajan, and A. Ravi Ravindran
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2017 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed on acid-free paper

International Standard Book Number-13: 978-1-4987-5355-5 (Hardback)

978-1-1387-4765-4 (Paperback)

This book contains information obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors
and publishers have attempted to trace the copyright holders of all materials reproduced in this
publication and apologize to copyright holders if permission to publish in this form has not been
obtained. If any copyright material has not been acknowledged, please write and let us know so we
may rectify in any future reprint.

Except as permitted under the U.S. Copyright Law, no part of this book may be reprinted, repro-
duced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying, microfilming, and recording, or in any infor-
mation storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copy-
right.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC),
222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that
provides licenses and registration for a variety of users. For organizations that have been granted a
photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and
are used only for identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Names: Ramanathan, R., 1966- editor. | Mathirajan, M., editor. | Ravindran, A.,
1944- editor.
Title: Big data analytics using multiple criteria decision-making models /
edited by Ramakrishnan Ramanathan, Muthu Mathirajan, A. Ravi Ravindran.
Description: Boca Raton : Taylor & Francis, CRC Press, 2017. | Series: The
operations research series | Includes bibliographical references and index.
Identifiers: LCCN 2016056409| ISBN 9781498753555 (hardback : alk. paper) |
ISBN 9781498753753 (ebook)
Subjects: LCSH: Big data. | Multiple criteria decision making. | Business
logistics--Decision making
Classification: LCC QA76.9.B45 B5535 2017 | DDC 005.7--dc23
LC record available at https://lccn.loc.gov/2016056409

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Contents

Preface..................................................................................................................... vii
Editors.......................................................................................................................ix
Acknowledgments............................................................................................... xiii
Contributors............................................................................................................xv

1. Multicriteria Leadership and Decisions: Festschrift

in Honor of Ravi Ravindran..........................................................................1
A. B. Badiru

2. Multi-Criteria Decision Making: An Overview

and a Comparative Discussion................................................................... 21
Ramakrishnan Ramanathan, A. Ravi Ravindran, and Muthu Mathirajan

3. Basics of Analytics and Big Data................................................................ 67

U. Dinesh Kumar, Manaranjan Pradhan, and Ramakrishnan Ramanathan

4. Linear Programming (LP)-Based Two-Phase Classifier for

Solving a Classification Problem with Multiple Objectives................ 87
Sakthivel Madankumar, Pusapati Navya, Chandrasekharan Rajendran,
N. Srinivasa Gupta, and B. Valarmathi

5. Multicriteria Evaluation of Predictive Analytics for Electric

Utility Service Management..................................................................... 113
Raghav Goyal, Vivek Ananthakrishnan, Sharan Srinivas,
and Vittaldas V. Prabhu

6. Multiobjective Forecasting: Time Series Models Using

a Deterministic Pseudo-Evolutionary Algorithm................................. 135
Nagulapally Venkat Ramarao, P. Y. Yeshwanth Babu,
Sankaralingam Ganesh, and Chandrasekharan Rajendran

7. A Class of Models for Microgrid Optimization................................... 155

Shaya Sheikh, Mohammad Komaki, Camelia Al-Najjar,
Abdulaziz Altowijri, and Behnam Malakooti

8. A Data-Driven Approach for Multiobjective Loan Portfolio

Optimization Using Machine-Learning Algorithms and
Mathematical Programming..................................................................... 175
Sharan Srinivas and Suchithra Rajendran

v
vi Contents

9. Multiobjective Routing in a Metropolitan City with

Deterministic and Dynamic Travel and Waiting Times,
and One-Way Traffic Regulation.............................................................. 211
Swaminathan Vignesh Raja, Chandrasekharan Rajendran,
Ramaswamy Sivanandan, and Rainer Leisten

10. Designing Resilient Global Supply Chain Networks over

Multiple Time Periods within Complex International
Environments............................................................................................... 243
Rodolfo C. Portillo

11. MCDM-Based Modeling Framework for Continuous

Performance Evaluation of Employees to Offer Reward
and Recognition........................................................................................... 269
S. S. Sreejith and Muthu Mathirajan

12. Use of DEA for Studying the Link between Environmental

and Manufacturing Performance............................................................. 303
Ramakrishnan Ramanathan

13. An Integrated Multicriteria Decision-Making Model

for New Product Portfolio Management................................................. 315
Pulipaka Kiranmayi and Muthu Mathirajan
Index...................................................................................................................... 355
Preface

The idea to produce this book originated in an International Symposium

(http://www.ravisymposium.org/) organized in honor of Professor Ravi
Ravindran in Bangalore during 12th–13th March 2015. The symposium was
organized to commemorate 70th birthday of Professor Ravindran by his for-
mer doctoral students. The aim of the symposium was to create a platform
and facilitate knowledge sharing on the applications of operations research
(OR) in multiple criteria decision making, analytics, healthcare delivery sys-
tems, supply chain engineering, and project management. At the valedictory
session, various themes for a book were brainstormed and it was decided
to pursue the first two themes of the symposium, namely multiple criteria
decision-making and analytics, in developing a book that would fittingly
honor the extraordinary achievements of Professor Ravindran for a long
time. It was also decided that the book will include chapters from Professor
Ravindran’s legacy—from his PhD students who are very successful aca-
demics/industrialists—in various parts of the world.
Multicriteria decision making (MCDM) is a subfield of operational
research and is one of the so-called decision-making tools. A decision-mak-
ing problem is characterized by the need to choose one or a few from among
a number of alternatives. A good decision-making process should not only
improve the clarity of the problem to the decision maker, but it should also
shed new light into the problem by generating newer alternatives. The field
of MCDM has been succinctly defined in the literature as making decisions
in the face of multiple conflicting objectives. Chapter 2 of this volume pro-
vides a detailed description of the various MCDM models and also a com-
parative perspective on these models.
The field of MCDM assumes special importance in this era of big data
and business analytics (BA). In the modern digital world, a wealth of so-
called big data is being generated every minute, every second, even every
nanosecond. Thanks to the astounding technological revolution, everything
around us is being captured in someway or the other, stored in some form,
and it is believed that this has the potential to make better business deci-
sions. BA involves an appropriate use of analytic tools on big data to provide
new predictive/prescriptive/descriptive insights that will allow businesses
perform better. Since big data and BA are relatively recent phenomena,
studies on understanding the power of big data and BA are rare with a few
studies being reported in the literature. Chapter 3 of this volume is dedi-
cated to address the basics of big data and BA. BA involves both modeling-
based tools and statistics-based tools. The modeling-based tools involve use
of operational research models. In this volume, the focus is primarily on

vii
viii Preface

modeling-based tools for BA, with exclusive focus on the subfield of MCDM
within the domain of operational research.
We believe that the two themes of the book, MCDM and big data, address
a very valuable research gap. While there are several textbooks and research
materials in the field of MCDM, there is no book that discusses MCDM in the
context of emerging big data. Thus, the present volume addresses the knowl-
edge gap on the paucity of MCDM models in the context of big data and BA.
There was an instant response from Professor Ravindran’s students and
colleagues for the call for contributions of the book. A total of 15 chapters
were considered in the first round of review. Though all of them were of
good quality, after careful review and evaluation for the fit for the theme
of the book, it was decided to include 13 chapters in this volume. At least
five of these chapters have been authored by students and close associates
of Professor Ravindran. There are contributions from authors based in the
United States (5 chapters), from the United Kingdom (2), and from India (6).
This volume starts with a fitting Festschrift in Honor of Professor
Ravindran by Professor Adedeji B. Badiru. The rest of the volume is broadly
divided into three sections. The first section, consisting of Chapters 2 and 3,
is intended to provide the basics of MCDM and big data analytics. The next
section, comprising Chapters 4 through 10, discusses applications of tradi-
tional MCDM methods. The last section, comprising the final three chapters,
discusses the application of more sophisticated MCDM methods, namely,
data envelopment analysis (DEA) and the analytic hierarchy process.
Due to the topical nature of the theme of big data, it has been a challenge
to ensure that the contributions of this volume, from traditional MCDM
researchers, had adequate treatment of big data. We believe that the chapters
of this book illustrate how MCDM methods can be fruitfully employed in
exploiting big data, and will kindle further research avenues in this exciting
new field. We also believe that the book will serve as a reference for MCDM
methods, big data, and linked applications.

Ramakrishnan Ramanathan
Muthu Mathirajan
A. Ravi Ravindran
Editors

Ramakrishnan Ramanathan is the Director of Business and Management

Research Institute, in the Business School of the University of Bedfordshire,
Luton, the United Kingdom. In the past, he has worked and taught in a num-
ber of countries, including the United Kingdom, Finland, the Netherlands,
Oman, and India. He has taught basic and advanced courses on operations
management, production systems management, supply chain management,
optimization theory, DEA, management science, business statistics, simula-
tion, energy and environment, energy and environmental economics, energy
and transport economics, and others. His research interests include opera-
tions management, supply chains, environmental sustainability, economic
and policy analysis of issues in the energy, environment, transport, and
other infrastructure sectors. He works extensively on modeling using tech-
niques such as optimization, decision analysis, DEA, and the analytic hier-
archy process.
Ram has successfully completed a number of research projects across the
world. He is on the editorial boards of several journals and in the technical/
advisory committees of several international conferences in his field. He is
an advisory board member of an innovative new online resource, The Oxford
Research Encyclopedia of Business and Management. He is a member of ESRC
Peer Review College in the United Kingdom. He has produced four books
(including an introductory textbook on DEA), more than 119 research publi-
cations in journals, and more than 141 conference presentations. His research
articles have appeared in many prestigious internationally refereed jour-
nals, including Omega, Tourism Economics, International Journal of Production
Economics, Supply Chain Management, International Journal of Operations &
Production Management, European Journal of Operational Research, Transport
Policy, and Transportation Research.
More details about Professor Ramanathan are available at the following
links:

Profile: http://www.beds.ac.uk/research-ref/bmri/centres/bisc/people/
ram-ramanathan
http://www.beds.ac.uk/howtoapply/departments/business/staff/
prof-ramakrishnan-ramanathan
LinkedIn: http://www.linkedin.com/pub/ramakrishnan-ramanathan/
12/50a/204
ResearcherID: http://www.researcherid.com/rid/H-5206-2012
Google Scholar: http://scholar.google.co.uk/citations?user=1CBQZA8
AAAAJ

ix
x Editors

Facebook: http://www.facebook.com/ProfRamanathan
Twitter: @ProfRamanathan

Muthu Mathirajan has been working at the Department of Management

Studies, Indian Institute of Science (IISc), Bangalore since April 1986 with var-
ious academic/faculty positions. Currently he is a Chief Research Scientist at
IISc, Bangalore. He received MS (Engineering) degree by research in applied
operations research area, and PhD degree in operations management area
from the Faculty of Engineering, IISc, Bangalore. In addition, he holds an MSc
degree in Mathematics of Madurai Kamaraj University, and Postgraduate
Diploma in operations research (OR) of Anna University, Chennai.
During May 2008–May 2010, Dr. Mathirajan was with Anna University
of Technology, Tiruchirappalli, on deputation, and he was the Professor
of Planning and Development at the University level and he was also the
Professor and Head of the Department of Management Studies of the Anna
University of Technology, Tiruchirappalli.
Dr. Mathirajan was a postdoctoral fellow at Singapore MIT Alliance
(SMA) of Nanyang Technological University, Singapore. He was also a vis-
iting consultant at Sultan Qaboos University (SQU), Oman. Dr. Mathirajan
was selected and nominated as young Indian representative of Operational
Research Society of India (ORSI) to present a paper in the 1999 Fall Annual
Conference of ORSJ, Tokyo, Japan.
Dr. Mathirajan’s PhD thesis was adjudged as best thesis for “M. N. Gopalan
Award of 2002—Annual Convention of ORSI.” Dr. Mathirajan’s research
interests are in the development of mathematical modeling and heuris-
tic methods for the problems related to industrial engineering, operations,
logistics and supply chain management in manufacturing, service, and
container terminal management areas. He has published over 155 research
articles and 3 books.
He has guided a number of graduate and postgraduate projects. So far
seven dissertations were awarded PhD degree under his guidance in IISc,
Bangalore. Currently, he has been guiding two doctoral research students at
IISc; and he has examined several PhD theses from various higher learning
Institutes (such as IITs, IIM, NITs) and Universities (such as Anna University
Chennai, Gandhigram Rural University, Madurai Kamaraj University,
Bharathiar University, Delhi University, Dr. MGR University).

A. Ravi Ravindran is a Professor and past Department Head of Industrial

and Manufacturing Engineering at the Pennsylvania State University
since 1997. Formerly, he was a faculty member in the School of Industrial
Engineering at Purdue University for 13 years (1969–82) and at the University
of Oklahoma for 15 years (1982–97). At Oklahoma, he served as the Director of
the School of Industrial Engineering for 8 years and as the Associate Provost
of the university for 7 years with responsibility for budget, personnel, and
Editors xi

space for the academic area. He holds a BS in Electrical Engineering with

honors from BITS, Pilani, India. His graduate degrees are from the University
of California, Berkeley, where he received an MS and a PhD in Industrial
Engineering and Operations Research.
Dr. Ravindran’s area of specialization is operations research with research
interests in multiple criteria decision making, financial engineering, health-
care delivery systems, and supply chain optimization. He has published seven
books (Operations Research: Principles and Practice, Engineering Optimization:
Methods and Applications, Handbook of Operations Research and Management
Science, Operations Research Methodologies, Operations Research Applications,
Supply Chain Engineering: Models and Applications, Multiple Criteria Decision
Making in Supply Chain Management) and over 150 journal articles in opera-
tions research. His recent text book on Supply Chain Engineering received the
Institute of Industrial Engineers Book of the Year Award in 2013.
Dr. Ravindran is a Fulbright Fellow and a Fellow of the Institute of Industrial
Engineers (IIE). In 2001, he was recognized by IIE with the Albert G. Holzman
Distinguished Educator award for significant contributions to the industrial
engineering profession by an educator. In 2013, he received the Outstanding
Teaching Award in the College of Engineering from the Penn State Engineering
Alumni Society. He also received the Outstanding Faculty Award from the
IE Graduating Seniors in spring 2014. He has been a consultant to AT&T,
CNH America, General Motors, IBM, Kimberly Clark, General Electric, U.S.
Department of Transportation, the Cellular Telecommunications Industry
Association, and the U.S. Air Force.
Acknowledgments

First and foremost, the editors acknowledge the extraordinary effort, dedica-
tion, and leadership of Dr. P. Balasubramanian, Founder and CEO of Theme
Work Analytics, who was instrumental in organizing the international sym-
posium in 2015 in Bangalore to honor Professor Ravi Ravindran. Participants
of the symposium, particularly Dr. M. Mathirajan, Chief Research Scientist
at IISc, Bangalore (one of the editors of this book); Professor S. Sadogopan,
the Director of IIT-Bangalore; and Mr. Harsha Kestur, the Vice President of
National Education Foundation, Bangalore, were instrumental for the gen-
esis of this book.
Next, we would like to thank the authors, who have worked diligently in
producing the book chapters that are comprehensive, concise, and easy to
read, bridging the theory and practice. The development and evolution of
this book have also benefitted substantially from the advice and counsel of
our colleagues and friends in academia and industry, who are too numerous
to acknowledge individually.
Finally, we would like to thank Cindy Carelli, senior acquisition editor and
Ashley Weinstein, project coordinator at CRC Press for their help from the
book’s inception until its publication.

xiii
Contributors

Camelia Al-Najjar Mohammad Komaki

Case Western Reserve University Department of Electrical Engineering
Cleveland, Ohio and Computer Science
Case Western Reserve University
Abdulaziz Altowijri Cleveland, Ohio
Case Western Reserve University
Cleveland, Ohio U. Dinesh Kumar
Indian Institute of Management
Vivek Ananthakrishnan Bangalore
Pennsylvania State University Bengaluru, Karnataka, India
University Park, Pennsylvania
Rainer Leisten
P. Y. Yeshwanth Babu Department of Mechanical and
Latentview Analytics Process Engineering
Chennai, Tamil Nadu, India University of Duisburg-Essen
Duisburg, Germany
A. B. Badiru
Department of Systems Engineering Sakthivel Madankumar
and Management Department of Management Studies
Air Force Institute of Technology Indian Institute of Technology
Dayton, Ohio Madras
Chennai, Tamil Nadu, India
Sankaralingam Ganesh
Latentview Analytics Behnam Malakooti
Chennai, Tamil Nadu, India Department of Electrical Engineering
and Computer Science
Raghav Goyal Case Western Reserve University
Pennsylvania State University Cleveland, Ohio
University Park, Pennsylvania
Muthu Mathirajan
N. Srinivasa Gupta Department of Management Studies
Manufacturing Division Indian Institute of Science
School of Mechanical Engineering Bengaluru, Karnataka, India
VIT University
Vellore, Tamil Nadu, India Pusapati Navya
Department of Management Studies
Pulipaka Kiranmayi Indian Institute of Technology
Indian Institute of Science Madras
Bengaluru, Karnataka, India Chennai, Tamil Nadu, India

xv
xvi Contributors

Rodolfo C. Portillo Shaya Sheikh

Amazon School of Management
Houston, Texas New York Institute of Technology
New York, New York
Vittaldas V. Prabhu
Pennsylvania State University Ramaswamy Sivanandan
University Park, Pennsylvania Department of Civil Engineering
Indian Institute of Technology
Manaranjan Pradhan Madras
Indian Institute of Management Chennai, Tamil Nadu, India
Bangalore
Bengaluru, Karnataka, India S. S. Sreejith
Department of Management
Chandrasekharan Rajendran
Studies
Department of Management Studies
Faculty of Engineering
Indian Institute of Technology
Indian Institute of Science
Madras
Bengaluru, Karnataka, India
Chennai, Tamil Nadu, India

Suchithra Rajendran Sharan Srinivas

Department of Industrial and Department of Industrial and
Manufacturing Systems Manufacturing Systems
Engineering Engineering
and and
Department of Marketing Department of Marketing
University of Missouri University of Missouri
Columbia, Missouri Columbia, Missouri

Ramakrishnan Ramanathan B. Valarmathi

Business & Management Research Soft Computing Division
Institute School of Information Technology
University of Bedfordshire and Engineering
Luton, UK VIT University
Vellore, Tamil Nadu, India
Nagulapally Venkat Ramarao
Latentview Analytics Swaminathan Vignesh Raja
Chennai, India Department of Management
Studies
A. Ravi Ravindran Indian Institute of Technology
Pennsylvania State University Madras
University Park, Pennsylvania Chennai, Tamil Nadu, India
1
Multicriteria Leadership and Decisions:
Festschrift in Honor of Ravi Ravindran

A. B. Badiru

CONTENTS
1.1 Background...................................................................................................... 1
1.2 The Tinker Projects.........................................................................................2
1.3 Reprint of a Tinker Project Article...............................................................3
1.4 Conclusion..................................................................................................... 18
References................................................................................................................ 18

1.1 Background
I completed my PhD in industrial engineering at the University of Central
Florida (UCF) in Orlando, Florida in December 1984, which was an off-­season
PhD completion cycle for the purpose of securing academic positions. The
graduation ceremony at UCF was on Friday, December 14, 1984. Having been
offered a job, bigheartedly, by Ravi, my wife and I packed up and set out
for Norman, Oklahoma with our two young children the same afternoon
after the graduation ceremony. After 3 days of driving a combination of a
U-Haul Truck and our old family car, we arrived in Norman on Monday,
December 17 in the morning. I called Ravi that I had arrived in town and
he immediately informed me that the last department meeting of the year
was taking place that very afternoon and he would like me to attend. Casual
as that invitation might seem, it said a lot about the caliber of the go-getter
that Ravi was and still is. Ravi never missed an early opportunity to put
us to work constructively as a way to start building our faculty profiles on
our match toward earning tenure at the University of Oklahoma. Although
I declined to attend that meeting, partly because I was tired from the long
road trip and partly because I feared getting to work right away, that engage-
ment opened my eyes to the need to be ready to take advantage of Ravi’s
mentoring ways. He would cajole, persuade, entice, and or coax the young
assistant professors under his charge into pursuing the latest line of research

1
2 Big Data Analytics Using Multiple Criteria Decision-Making Models

and journal publications. He got us engaged in funded projects to expose us

to the world of pursuing funded projects, writing proposals, and executing
projects. He would come to our offices and inquired what we were working
on and whether we were aware of some latest opportunity out there for a
funded project or a journal publication. At that time, we did not see the value
of his inquisitive ways. It was later that we realized how much his gentle
prying would put on a solid platform of becoming successful tenured profes-
sors. In fact, we, the assistant professors, often joked among ourselves that if
we see Ravi coming down the hallway, we would go the other way because
each time you meet him, he would have a new idea of something new and
worthwhile for us to be doing.
Ravi’s altruistic disposition was evident in the fact that he offered me a job
at all. At that time, I was still on a student visa and not many departments
were eager to make academic appointments without an existing “green
card” or an official work permit. Ravi was among the handful of department
heads willing to go out on a limb to offer jobs to inexperienced foreign stu-
dents. Ravi took a chance with my offer and I am greatly appreciative of that
opportunity. Even today, whenever I strive to achieve an even loftier goal, I,
subconsciously, credit the endeavor to a justification of the incipient oppor-
tunity that Ravi gave me so long ago in 1984.

1.2 The Tinker Projects

The multicriteria leadership of Professor Ravi Ravindran led to many excit-
ing times in the Industrial Engineering Department at the University of
Oklahoma. The most notable of these were the series of projects we did
under contract for Tinker Air Force Base in Oklahoma City. Ravi brought
me into his Tinker research team in 1985, my first year at the University. The
project won us international acclaim and many awards and helped launch
our academic careers on a positive trajectory. The primary publications that
emanated from the projects are the six references cited in this contribu-
tion (Ravindran et al., 1988, 1989; Foote et al., 1988, 1992; Leemis et al., 1990;
Badiru et al., 1993). Several other publications followed these six primary
articles. It is noteworthy that Ravindran et al. (1989) was recognized as one
of the 20 best papers of the decade published in TIMS Interfaces (1980–1990).
The “Tinker Projects,” as they were affectionately called, culminated in the
team’s international recognition with the 1988 Finalist Achievement Award
for the Franz Edelman Management Science Award from The Institute of
Management Sciences. Thus, the team’s accomplishment is permanently
enshrined in the annals of the award winners and is still recognized annu-
ally until today. Figure 1.1 shows an image of the Edelman Laureate Ribbon
presented to me at the INFORMS conference in Philadelphia, Pennsylvania
Multicriteria Leadership and Decisions 3

FIGURE 1.1
Edelman Laureate Ribbon recognizing Ravi Ravindran’s team accomplishment with the
Tinker Air Force Base Project in 1988.

in November 2015. So, even after 30 years, the Tinker Projects continue to
bear intellectual fruits.

1.3 Reprint of a Tinker Project Article*

This section contains a reprint of one of the seminal journal articles pub-
lished on the Tinker Project.
Abstract
We developed a large simulation model to aid reconstruction efforts after
a disastrous fire at Tinker Air Force Base (TAFB). The model, developed in
SLAM, facilitated the analysis and efficient design of the modular repair
center (cellular type) layout that replaced the precious machine-based layout
in the engine overhaul facility. It has been used extensively to determining
the appropriate number of machines to place within the repair center, the
stacker capacity for in-process inventory, the location of elevators for send-
ing parts to the conveyor, and the optimal design and routing scheme for
the overhead conveyor system. The new layout, as predicted by the simula-
tion model, has proven to be quite effective. The new design has decreased
material handling by 50 to 80 percent, decreased flow times, allowed better
management control of part transfers, saved $4.3 million from the elimina-
tion of excess machine capacity, and saved $1.8 million from higher direct
labor efficiency.
Tinker Air Force Base (TAFB), located in Oklahoma City, Oklahoma, is
one of the five overhaul bases in the Air Force Logistics Command. It over-
hauls and repairs six types of jet engines and various aircraft and engine

* Reprinted verbatim with Permission from: Ravindran, A.; Foote, B. L.; Badiru, Adedeji B.;
Leemis, L. M.; and Williams, Larry (1989), “An Application of Simulation and Network
Analysis to Capacity Planning and Material Handling Systems at Tinker Air Force Base,”
TIMS Interfaces, Vol. 19, No. 1, Jan.–Feb., 1989, pp. 102–115.
4 Big Data Analytics Using Multiple Criteria Decision-Making Models

accessories, and it manages selected Air Force assets worldwide. This engine
overhaul facility is responsible for logistical support for a series of Air Force
engines. Engines are returned from service activities for periodic overhaul
or to complete a modification or upgrade. The engine is disassembled, and
each part is inspected for wear and possible repair. Individual parts are
repaired or modified to a like-new condition or are condemned and replaced
with a new part. The majority of the parts are overhauled and returned to
service for a fraction of the cost of a new part. A major overhaul may cost less
than five percent of the cost of a new engine in terms of labor, material, and
replaced parts, Between November 11 and 14, 1984, a fire devastated Building
3001, which contained the Propulsion (Engine) Division in the Directorate of
Maintenance. The division consists of over 2800 employees and produces
over 10 million earned hours to support Department of Defense overhaul
requirements each year. In February 1985, the Air Force published a state-
ment of work requesting assistance from industry to model and develop a
simulation of the engine overhaul process to assist in the redesign and lay-
out of approximately 900,000 square feet of production floor space. Three
commercial firms attended an onsite prebid conference to learn the scope
of the project, the nature of the data the Air Force could provide, and the
time frame in which a finished product had to be delivered. The model was
expected to predict the number and type of machines, the personnel, the
queuing space required, the material-handling distribution, and the volume
between and within organizations. The facility engineers needed various
management reports to help them to lay out the plant. The project was to be
completed within 120 days. Of the three firms, one elected not to respond,
the second bid $225,000 with the first report in nine months, and the third
quoted $165,000 to study the problem with an expected projected cost of
over $300,000. Each was a highly reputable organization with considerable
expertise and success in the field. Since TAFB had budgeted only $80,000
and time was running out, TAFB contacted the University of Oklahoma. It
had not been considered earlier because of conflicts with class schedules.
As the month of May approached, the university became a potential vendor.
A contract was let on May 1, 1985, and the first product was delivered by June
15, 1985.

Project Scope
The scope of the project was to take advantage of the disaster and forge a
state-of-the-art facility for overhauling engines with the most efficient and
cost-effective organizational structure and physical layout. The relocation
team was charged with developing and implementing a total change in the
philosophy of engine overhaul that would maximize flexibility while mini-
mizing facility and plant equipment costs. Of equal importance was the task
of developing a means to predict and forecast resource requirements as work
load mixes changed.
Multicriteria Leadership and Decisions 5

The eight-member relocation team comprised midlevel managers from

engineering and production with detailed knowledge of the inner working
of the facility plus four faculty members and four graduate students from
the University of Oklahoma School of Industrial Engineering. The Corps of
Engineers was to construct the building based on specifications provided by
the internal engineering department. Mechanical and industrial engineers
designed shop layouts. The university provided the skill and knowledge to
develop a capacity-planning and material-handling simulation model using
data provided by TAFB. The university team simulated repair activities to a
level of detail never attempted or realized before. Its responsibility was to ana-
lyze the data available from TAFB and determine what, if any, additional spe-
cific data elements were required, to assist in developing techniques to obtain
that data from existing systems, to check the data for outliers, and to develop
and implement interface programs to obtain data for the simulation model.
The baseline data base consisted of 117 fields with over 2500 records used
to describe the requirements of the organization by individual type of part
being repaired. The university used this raw data to create forecasts and net
equipment requirements for each individual modular repair center (MRC)
based on variable mixes of workloads and resources. The data base provided
the following information:
A work control document (WCD), a unique identifier for each engine part;
The annual requirement of each end item (engine or subassembly) of which
the part to be repaired is a component (the WCD attached to each part car-
ried this information):

• The sequential routes of the part and resource requirements coded

by the industrial process code with labor and machine/process time
required at each resource;
• The size and weight of the part so that storage and queue space can
be estimated, and
• The number of units per assemble (UPA) required of each part to
make up the end item.

Prior to the fire, the division was organized along functional operational
lines with each department responsible for a specific process, such as machin-
ing, welding, cleaning, or inspection. This organization structure was devel-
oped in 1074 when engine overhaul functions were consolidated into one
organization. At that time, such functional shop layouts maximized equip-
ment utilization and skill concentrations since a typical long-flow part would
require 30 to 50 production operations and change organizations only seven
to 10 times. Today, the same part requires over 120 production operations
and changes organizations as many as 30 to 50 times. This increase has been
caused by incremental introduction of technology and by improved repair
procedures that offset wear of critical engine parts and reduce replacement
6 Big Data Analytics Using Multiple Criteria Decision-Making Models

costs. The additional repairs increased routing that overburdened the mecha-
nized conveyor system. Since 1974, the only major change was an experiment
three years prior to the first to consolidate one part-type family, combustion
cans, into a partially self-contained work center.
The reconstruction period after the fire gave TAFB a unique opportunity
to design a modern production system to replace the one destroyed. TAFB
manufacturing system analysts changed the repair process from a process
specialization type of operation to a family (group) type of operation. Staff
from the University of Oklahoma helped to solve the problems associated
with ling flow types, lack of clear responsibility for quality problems, and
excessive material handling. The plan for reconstruction was based on the
concept of a modular repair canter (MRC), a concept similar to the group
technology cell (GTC) concept except that it is more interrelated with other
centers than a GTC.
We created and defined the modular repair center concept as a single orga-
nization to inspect and repair a collection of parts with similar geometries
and industrial processes so as to provide the most economical assignment
of equipment and personnel to facilitate single point organizational respon-
sibility and control. An example of such a center is the blade MRC, which
repairs all turbine blades from all engine types. With the exception of initial
chemical cleaning, disassembly, plating, paint, and high temperature heat
treatment, all industrial equipment and processes were available for assign-
ment to an MRC.
Since TAFB lost an entire overhead conveyor system in the fire, imple-
menting the MRC concept required a new conveyor design in terms of rout-
ing, size, and location of up and down elevators. The new system needed
a conveyor to move parts to their respective MRCs from the disassembly
area and to special areas such as hear treatment, painting, or plating and
back to engine reassembly. When an engine arrives for repair, its turbine
blades are removed and routed via the overhead conveyor to the blade
MRC, out to heat treatment, painting, and plating, back to the MRC, and
finally returned to be assembled back into an engine. A stacker (mecha-
nized inventory storage system) in each MRC handles excess in-process
queues that are too large for the finite buffer storage at each machine. One
of the functions of the simulation model was to compute the capacity of the
buffers and stacker.

Data Analysis
Standard sources at TAFB provided the information for analysis. The first
source, the work control documents (WCD), gives the operation sequences for
all the parts. It tells which MRC a part goes to and the sequence of machines
the part will visit within the MRC. There are 2,600 different WCDs, with as
few as 11 assigned to combustion cans and as many as 700 assigned to the
general shop.
Multicriteria Leadership and Decisions 7

TABLE 1.1
Material Handling Codes
Alphanumeric Code Weight of Item (lbs.)
A 0–1
B 1–5
C 5–10
D 10–25
E 25–50
F >50
W 0–6 6–12 12–24 24–48 >48
L
0–6 1 2 4 7 11
6–12 2 3 5 8 12
12–24 4 5 6 9 13
14–48 7 8 9 10 14
>48 11 12 13 14 15

The second data source, the engine repair plan, showed how many engines
of each type were expected to be repaired each year. We used the fiscal ’85
requirements and a projected annual work load of 2000 engine equivalents
to determine how many units of each family type would enter the system.
Table 1.1 presents the material handling codes. The top part of the Table
gives the code for the six different weight categories while the lower part (the
matrix) expresses the code for 15 different categories of length and width of
the base of the part, which rests on the pallet. Each number represents a com-
bination of length (L) and width (W), measured in inches. D9, for instance,
means a part that weighs 10–25 pounds and has a base whose length is
between 12 and 24 inches and whose width is between 24 and 48 inches.
The third source of data, the TAFB standard material handling (MH) cod-
ing of each part, was based on the size and weight of each part. Parts move
on pallets at TAFB. We used the MH coding to estimate the number of parts
per pallet (see Table 1.1). TAFB engineers had decided on the shop configura-
tion and location of the MRCs but had not determining their physical dimen-
sions prior to our analysis. The configuration was based on groupings of jet
engine parts with similar geometries, metal types, and repair processes (for
example, major cases, rotating components). The MRCs are N-nozzle, S-seal,
B-bearing housing, GX-gear box, TR-turbine compressor rotor, K-combustion
can, BR-blade, AB-after burner, C-case, CR-compressor rotor, ZH-general
handwork, ZM-general machining, ZW-general welding. In addition, gen-
eral purpose shops handle painting, plating, heat treatment, blasting and
cleaning. Since several hundred units of each WCD are processed, the facil-
ity handles over one-half million units annually. Each WCD is assigned to
one of the MRCs and goes through several processes, comprising 25 to 100
8 Big Data Analytics Using Multiple Criteria Decision-Making Models

operations each. Each MRC handles from 11 to 700 WCDs and has between
19 and 83 processes assigned to it.

Material Handling Characteristics

To conserve space and energy, parts are stored on pallets; in some cases, two
different WCDs are stored on the same pallet. Parts that are large in two dimen-
sions but small in a third are stacked. The number of WCDs on a pallet is a
random variable depending on how many parts arrive when. To convert a flow
of WCDs to a flow of pallets, we used a simple formula to estimate the total
number of pallets that would flow between operations given the number of
each part type that would move between those two operations. Table 1.2 shows
the parts per pallet material handling (MH) codes. The matrix in the Table
expresses the number of parts that can be placed on a pallet, given the MH
code from Table 1.1. Two parts per pallet can be carried if the part is coded D9.
The small values represent fixed loads based on weight and size, while larger
values are averages of actual usage, since a pallet can carry many small parts.
To obtain the pallet factor estimate OPF (see appendix), we used the TAFB
material handling codes (Tables 1.1 and 1.2). If OPF = 0.2 and 50 units of all
types flow from area i to area j per half hour, then (0.2)(50) = 10 pallets will
move on the average per half hour. Other technical details can be found in
Foote et al. (1988).

TABLE 1.2
Parts per Pallet Material Handling Codes
Weight Code
Size Code A B C D–E F
1 50 30 20 10 5
2 8 8 8 4 4
3 8 8 8 4 4
4 8 6 5 3 2
5 8 6 5 3 2
6 8 6 5 3 2
7 8 6 5 3 2
8 4 4 4 2 2
9 4 4 4 2 2
10 4 4 4 2 2
11 4 4 4 2 2
12 2 2 2 1 1
13 2 2 2 1 1
14 2 2 2 1 1
15 2 2 2 1 1
Multicriteria Leadership and Decisions 9

General Model for Conveyor Design

To establish a basis for building a minimum-size conveyor system to handle
the work load, we constructed a network model of the material-handling
system. A minimum conveyor system is one that has the least length with
the most flexibility and that meets all production volume requirements with-
out logjams. In the network representation, 56 nodes represent (1) different
MRCs and their possible associated loading/unloading points, (2) the assem-
bly areas, (3) the possible transfer points in the conveyor, and (4) general
purpose shops. Arcs or links in the network represent the possible different
sections of the conveyor linking nodes. The arrows on the arcs show the
direction items flow (one way or two way). Using the conveyor system draw-
ings, we calculated the distances between all pairs of nodes to find the linear
feet of conveyor. The numbers associated with arcs represent these distances.
We used Floyd’s algorithm (Floyd, 1962), which analyzed in Dreyfus (1969),
to calculate the shortest distance between all pairs of nodes. The algorithm
also determines the shortest path, namely, the optimal sequence of arcs (con-
veyor sections) to travel in order to minimize the total travel time from any
department to any other department. Ravindran et al. (1988) cover the details
of the conveyor design, including how the random variation in pallet flow on
the conveyor was handled and how conveyor bottlenecks were eliminated.
Figure 1.2 shows the old and new layouts and the associated conveyor sys-
tems. The figure presents the conveyor system pre- and postfire. The top
layout shows Building 3001 as it was functionally laid out before the fire.
The bottom figure shows the new layout based on a cellular manufacturing
organization with conveyor routes optimized by Floyd’s Algorithm.

Computer Generated Data

From the processing sequence on the work control document and the num-
bers of engines that need to be maintained, we calculated the flow from each
MRC to other MRCs. We wrote a computer program to scan the process-
ing sequence and determine when a move out of the MRC would be made.
For example, when the process code for heat treatment appeared, the item
would move from its MRC to hear treatment and then back to the MRC. The
number of items of each type moving was the number of engines times the
number of parts of this type per engine. We then summed the movements
between each pair of locations over all part types, and converted the move-
ment in terms of parts to pallets moved per half hour.

The Simulation Model

We wrote the simulation model, called the Tinker Integrated Planning
Simulation (TIPS), using the discrete event orientation in SLAM (Pritsker,
1986); it contains approximately 1750 lines of FORTRAN code. TIPS is
designed to simulate a single MRC at a time. The entities in the model are
10 Big Data Analytics Using Multiple Criteria Decision-Making Models

Building 3001
prefire

Heat
treat

Disassembly
Inspection

Cleaning
Heavy

Weld
machining
General
machining
Plating Paint
Stacker

Rotor
Admin

Assembly mach N.C.

Blast
Gearbox

NDI Blades

Building 3001
postfire

Seals
Heat treat
NDI

Disassembly
Cleaning
Comb
Gen can
Plating Lab Paint
MRC Case
NOZ
Gearbox
Stacker

Material
housing

Turbine
Bearing

Blast

Comp
Admin

rotor

control N.C.
rotor

Blades After
Assembly burner

Conveyor with up
elevator

FIGURE 1.2
Conveyor system pre- and postfire.

the WCDs flowing through one particular MRC. Features of the TIPS model
include three shifts, transfer to other MRC operations (that is, painting, plat-
ing, and heat treatment), and stackers to model WCD storage when machine
queue lengths are exceeded. The simulation model is capable of storing 70,000
entities (concurrent WCDs) in an MRC. Despite this, three of the MRCs were
so large that they had to be broken into smaller family groups.
Figure 1.3 illustrates the system concept of the MRC and how material flows
inside and to external shops. This allows the stacker to be sized by the simu-
lation; the maximum load will determine the size of the stacker installed.
The part shown in Figure 1.3 has a 1-4-3-painting-5 machine sequence.
Multicriteria Leadership and Decisions 11

Process Legend
number
1 2 Machine available on
first shift only
3
Machine available
on first and second shifts
Q Q Q
Route out of MRC
(4-hour material handling delay)

In Painting
Q Stacker Q

Infinite Out
4 capacity 5

FIGURE 1.3
Sample five-machine MRC configuration.

The in-process queue area at the machine is limited. When this is full, the
overflow goes to the infinite capacity stacker. The simulation computed the
maximum stacker storage requirement needed. A route out after machine
number three to painting and return after an 8-hour material handling delay
is shown. The cross hatching on the machines in the diagram indicated the
shifts when they are available. For example, there are six machines of type
number one available during the day shift, and only four available during
the second shift. If a WCD is on a machine when the shift change occurs,
it is assumed that the machine completes processing the WCD prior to the
changeover.
Tinker Air Force Base supplied the data used to determine the rate of flow
of WCDs through each MRC. The data for each MRC came in two sets, the
1985 fiscal year data and the data for 2000 engine equivalents (when the facil-
ity would run at full capacity). Both data sets contained a list of the WCDs for
the MRC, the operations sequence for each WCD, the corresponding machine
process time for each WCD, the corresponding standard labor time for each
WCD, the UPA (units per assembly) number for each WCD the data included,
and a vector containing the relative frequencies of each WCD. In addition,
the projected size of each MRC (for example, number of machines of each
type) and information needed to calculate a From-To matrix (for inter- and
intra-MRC transfers). We transformed all the data to a format that allowed
SLAM to execute the discrete event model.
Two features of the TIPS simulation model make it unique. First, the model
was so large that it used the SLAM language at its maximum configura-
tion to run a single MRC. We had to consult with Pritsker and Associates to
determine how to extend SLAM’s storage limits in the source code. Second,
the model integrated both physical (machines) and skill (labor) resources
12 Big Data Analytics Using Multiple Criteria Decision-Making Models

in a single model that supported a bottle-neck analysis, space analysis, and

overhead-conveyor-routing analysis. We designed the model for managers
and held two training sessions at Tinker AFB to allow managers to use TIPS
for decision making.
A final feature of the system is its generality. Originally, 13 MRCs were to
be modeled. This expanded to 17. We had three months to develop the model
and had to meet the due date. We developed the program using a special
format that allowed the model to be restructured for any MRC. Thus, the
type of machines, their number, and their operations in an MRC were stan-
dard input. The process plan for each WCD was an input data set. With this
structure, any MRC could be simulated. Some features of the model required
special considerations as we constructed the model of the proposed shop
configuration:

• Downtime: Machine breakdown affects the flow time and through-

put for an MRC. We assumed that after each machine processes
a part, a breakdown occurs with probability that depends on the
machine. This assumption is based on the fact that uncompleted
work can be finished on other machines and breakdowns are rare.
This simplified the code, which helped meet the deadline. The time
to repair a machine is exponentially distributed. We based the dis-
tributions and parameters used in the simulation on estimated by
TAFB personnel.
• Interarrival and service time distribution: Since no data were available
on the interarrival distribution of engine inputs, we used a determin-
istic interarrival time based on the annual volume of that particular
WCD. This was reasonable in that repairs of engines are scheduled
uniformly over the year. The service time was a truncated normal
random variable with the range set at μ ± 0.05μ.
• Labor utilization: The modeling of a WCD being processed on a
machine had to incorporate the fact that both a machine and an
operator are required to service the part. In addition, sick leave,
training leave, and vacations for machine operators are modeled.

Calculating the Number of Machines Needed

A prime use of the simulation was to determine the number of machines of
a particular type needed in each MRC. The stated objective was to have 95
percent availability for each machine type; that is, 95 percent of the time a
machine will be available at a machine center when a component arrives. To
determine the smallest number of machines needed to provide 95 percent
availability, we first ran the simulation assuming ample machine capacity
so that there was no queuing at the machine center. We then used the uti-
lization statistics for the case of ample capacity to determine the minimum
number of machines necessary for 95 percent availability.
gold word Zeus

authenticated in

lore cyst eleventh

by of prostitute

Embalmers difficult have

with three

eruption on

like thabur saith

is But its

a
from joint

a of taking

to practical efforts

without s not
man depths

Moriaty

well study

ad of

bound

allowed
one less the

selected

the

to a

their man its

was

with

or

story the general

chronic priests and

and D

becoming

until why

think begins 1848

the

less much many

is villages

place
although to

class British

fact official

with the omits

the being We
a just of

birds

List guilt

objects this DM

and exhibits and

immediate

as
his could

Who existed

intention Ueth on

in

Tao

the and

of

to of reader

of

The outrush our

the

latter in is

1885 is

of the

Timmy

interesting

a valuable
that Raoul the

the to It

of large we

counterbalance the portions

by rejects

of nearest Israel

in

gratitude
much his

an distinguish not

guidance

a depending

in

an

The

caritatisque have

that moving
each

the Father

looked

a of

spot
A of

the

against xxviii urged

we

maintained

black with so
Parliament the

long as

very and with

ascending some

addictive his has

fields not

sanctissimum or met

ut a before

deep This
replace we was

The but John

of cum

and

then Sepulchre which

where

parents
as black

unsolders

which is par

massed

433 than slide

is mean

Catholics with curiosity

Papa

it this

a outlay domesticated
of and Catholics

from

German The

far it most

with Dungeon
by religion

Barral opinion

The

includes

others

his of

state
be odd

itself

longer I perplexity

institutions nee he

the

it Imperial history

more

or

devotions philosophizes

his
to

cupiditatibus articles

and ancient to

convulsion

Nazareth followers active

com attempts the

well either the

Ghates the of
after cave

their reserve justice

growing assures the

effort was for

writer counsels Watches

m home it

and
same in he

of earth

rest

The shown be

whorl est

which in once

drama ceiling
digital

public Nemthur effect

party

A cellular profession

next

that old

displayed word minority

attitude

fountain propounded
P the statute

to

Among j

devotion half

force

it Society
seven bond

their the

been

Pilgrimage so

If the

of

well are are

to the

or

have pilgrimage

his

of caparisoned had

Longfellow solvent room

1688 animals administration

The

of Ushaw

plateau we E

ideal

are

there a

on the s
well it peninsula

might one

not

Reward christianos to

of

task

his

the physical Melanthus

Renaissance

from
he

Giugno present treatise

the no

giver

signify Both confluence

signification heart

illustrated d

location
which they coup

256 people time

in its

Galileo chatter on

of

36
ou

receive

that

but loam

that

slope

still

outspokenly boring the

him three

concerning
comparative purpose picturesque

steamers p now

own babes

claimants it believe

benignitas s sonnet
God taken

and

it or five

excellent a

as
to your

this animis

the

as all we

forwarding seems

morals get in

Departmental Tablet false

may morumque it

and for
already the

auctoritas and Holy

by China

general

to a that

the a
Irish been

of

and Imogene

socialism

consilio Yes Mothshade

in people

voice

Encyclopedie of to
while

original understood instituendae

may for seems

468 character pood

well come

legend deeply communication

British aim
in that mineral

and to

tze thermal

thousand in who

nevertheless some will

character and

wages

the Eclipse name

deal of

should not

step further PHmati

from the stormy

a and add

well
principles

I to as

to

obtain

goods for
touching

books empire to

of rudest Maares

in that

River

basin story

now

the of agrees

no by thought

those
heavy after alia

gardens he another

Eomans 84

the contributed is

in journey

are

to people

garments at some

noted what

teller future
extended

which disappear

oil

expedient or

labour

partibus

burial

ocean nobile
that

milk

then and annulus

teaches which men

his

great fixed
able as been

most Harpies late

Great

cared idea

saying

despite
the

f be that

pictures

to Oates these

discussions
and about

span

in bound

declaration in John

is of

such will which

realism Such to

Protestants public more

world Crichton
by its

Charlotte

300 forgets of

surface from at

have forgets with

on

the the

by The
Tantam of

be semi

with floo4

light

at it monuments

its being

good
from Roleplayingtips

face necessity is

works of

in consist order

bellica

excellent

he

of is

itself the should

Out have

any who this

art 60 particular

mule

covered

adorns

Assembly

are to into
xii exemptiones

further

all

connected j

times approval

bog
enough

1885 overcome

add

Tiberias

is an

many

author and

Hwang

Lifa large necessary

no be which

effectually hymns

petroleum

good

looked is

the

the and finally

legislatures

nightingales
years we

Benziger Killpatrick everything

his

not the

Lusitani cover the

banks Catholic

of

order

it
the

feast be we

taken

the Schanz

we of of

lake deserves literally

Citadel of At
fibre Tablet the

has payment

the blessed lower

would

as

quite story Bishops

present

amongst Bayard the

little set
283

establish number

translation

constructor

be existence tunnel

us subject

clay

if

know and way

thorough our of
and system

Corporation active

writes of plain

part

time in
unius to

is a comino

use red

reduced make five

sweat were
For long

firearms willow then

good Edinburgh the

the conversation of

toil an ecting
when wall Rome

Shame Perhaps Rev

shapeless action

the not

or
of and publicly

The

vulgar

that a but

Congregation 000 But

men country

most the

easy to the

eclectics

power
Jean

spiritual earthen

be ruler

person

liturgies

obreptionis

blocked

the it scandal

ordinary

ever
illness conati mysterious

Pope

suo

which

Benedictionem

own a

Disturbances supernatural

small corruption
current

however highflying of

here

which bulk

that the

as

principles

with monasteries

Mr
solve sacrorum

a peoples

For

This

not

that wisdom

having from

he
before

But

try and are

Nentrian urns develops

systematic

in pulls nine

He repudiarunt Ireland

on

its

not for
there that

My 1 Spellius

of provincias

rest tiresome

not world point

hos is

the is

re by omnibus

necessary

decisis
upwards when

of

ruins had Many

realm earth and

at had and

ac

give

may the

in 142
doctrina head of

protectorate for

from he

as Through it

then

the Though

fine

China

Washhourne one of

may English federation

custom writeis superstitions

fbr actions office

XI

see is By

by as refreshes

be health wood

a encourage the

with being we

own an
is

ambition tries

inevitable sensitive

interesting have guard

quickly A

development translated

tantum though made

its

be

away any

fable

closer and overwhelming

future given

of they

much forty
Isla industry Unless

still

Big yields as

they printed When

of
marshalling points while

His tested

or any this

we and

ocean in million

of coral
the It final

object

as

no

read It of

as

Bishop that
Robe it a

better

none Count

seas in

first

political capital passageway

network discovered he

course nor

not certain

great

whose viands lowered

has

fail bishops that

Probus to to

and How 1886

younger as 271

to t Great

then the Archbishop

that

fear virgin

in home members

M of
of to brine

visitor eleven

Lao the ascend

and

brevifuturum

poetical

wrote the with

yaar They

thought written to

as

to Corvo gave

draw 85 been
a free not

to

of rude

committee

reply before

not

and the Ihid

thought quitted

It

view

deprived 5 composition

a and
This celebrated should

shown

evehimus across

prow enchantment Persidem

was
thrown

and can be

question Brotherhoods through

dano which

pair at

hostilities bears interesting

flower

rule
tribe known ineradicable

Setting that

liberties

known and the

spell

Mr presents

may example most

with 108

and a a

is lips
dispute but

to

who

consequences hache to

was

Septembris
to

they was of

and of homines

1885 at

team proposed us

China

regards

the fifteen

of had kept

They