Professional Documents
Culture Documents
Textbook Grouping Genetic Algorithms Advances and Applications 1St Edition Michael Mutingi Ebook All Chapter PDF
Textbook Grouping Genetic Algorithms Advances and Applications 1St Edition Michael Mutingi Ebook All Chapter PDF
https://textbookfull.com/product/advances-in-metaheuristics-
algorithms-methods-and-applications-erik-cuevas/
https://textbookfull.com/product/advances-and-applications-of-
optimised-algorithms-in-image-processing-1st-edition-diego-oliva/
https://textbookfull.com/product/genetic-algorithms-in-java-
basics-jacobson-lee-kanber-burak/
https://textbookfull.com/product/genetic-algorithms-and-machine-
learning-for-programmers-create-ai-models-and-evolve-
solutions-1st-edition-frances-buontempo/
Low Power Wireless Receivers for IoT Applications with
Multi-band Calibration Algorithms 1st Edition Michael
W. Rawlins
https://textbookfull.com/product/low-power-wireless-receivers-
for-iot-applications-with-multi-band-calibration-algorithms-1st-
edition-michael-w-rawlins/
https://textbookfull.com/product/preimplantation-genetic-testing-
recent-advances-in-reproductive-medicine-1st-edition-darren-k-
griffin-editor/
https://textbookfull.com/product/digital-media-steganography-
principles-algorithms-and-advances-1st-edition-mahmoud-
hassaballah-editor/
https://textbookfull.com/product/swarm-intelligence-algorithms-
modifications-and-applications-1st-edition-adam-slowik/
https://textbookfull.com/product/data-mining-algorithms-in-c-
data-patterns-and-algorithms-for-modern-applications-1st-edition-
timothy-masters/
Studies in Computational Intelligence 666
Michael Mutingi
Charles Mbohwa
Grouping
Genetic
Algorithms
Advances and Applications
Studies in Computational Intelligence
Volume 666
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
About this Series
123
Michael Mutingi Charles Mbohwa
Faculty of Engineering Faculty of Engineering and the Built
Namibia University of Science and Environment
Technology University of Johannesburg
Windhoek Johannesburg
Namibia South Africa
and
Recent research trends have shown that industry is inundated with grouping
problems that require efficient computational algorithms for grouping system
entities based on specific guiding criteria. Grouping problems commonplace in
industry include vehicle routing, container loading, equal piles problem,
machine-part cell formation, cutting stock problem, job shop scheduling, assembly
line balancing, and task assignment. These problems have a group structure with
identifiable characteristic features, that is, the need to form efficient groups of
entities according to guiding criteria, and the need to allocate those groups to
specific assignees in order to satisfy the desired objectives. It is interesting to note
that, across all the spectrum of these problems, grouping and allocation criteria are
inherently very similar in nature.
The wide spectrum of real-world grouping problems, the striking similarities
between their features, and the multi-criteria decisions involved are three major
motivating factors behind the research momentum in this area. However, more
challenging issues in this field have appeared in recent researches. First, there is an
ever-growing need to address uncertainties in various grouping problem situations.
Second, decision analysts in the field often call for multi-criteria decision approa-
ches by which multiple criteria can be handled simultaneously. Third, researchers
and decision analysts have realized the need for interactive, population-based
algorithms that can provide alternative solutions rather than prescribe a single
solution to the decision maker. Examples of such approaches are tabu search,
particle swarm optimization, ant colony optimization, simulated evolution algo-
rithm, simulated metamorphosis algorithm, genetic algorithms, and grouping
genetic algorithms. Thus, in sum, recent research has emphasized the need for
development of interactive multi-criteria computational algorithms that can address
grouping problems, even in uncertain or fuzzy environments.
Evidently, notable research has focused on advances in genetic algorithms and
related hybrid approaches, with application in various problem areas. Current
research trends tend to show that there is a high potential for remarkable advances
in genetic algorithms and its variants, specifically in grouping genetic algorithms.
Genetic algorithm-based approaches offer a more user-friendly, flexible, and
vii
viii Preface
Part I Introduction
1 Exploring Grouping Problems in Industry . . . . . . . . . . . . . . . . . . . . 3
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Identifying Grouping Problems in Industry . . . . . . . . . . . . . . . . . 5
1.2.1 Cell Formation in Manufacturing Systems . . . . . . . . . . . 5
1.2.2 Assembly Line Balancing . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.3 Job Shop Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2.4 Vehicle Routing Problem . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.5 Home Healthcare Worker Scheduling . . . . . . . . . . . . . . 10
1.2.6 Bin Packing Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.7 Task Assignment Problem . . . . . . . . . . . . . . . . . . . . . . . 13
1.2.8 Modular Product Design . . . . . . . . . . . . . . . . . . . . . . . . 14
1.2.9 Group Maintenance Planning. . . . . . . . . . . . . . . . . . . . . 15
1.2.10 Order Batching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.2.11 Team Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.2.12 Earnings Management . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.2.13 Economies of Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.2.14 Timetabling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.2.15 Student Grouping for Cooperative Learning . . . . . . . . . 22
1.2.16 Other Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.3 Extant Modeling Approaches to Grouping Problems . . . . . . . . . 23
1.4 Structure of the Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2 Complicating Features in Industrial Grouping Problems . . . . . . . . . 31
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2 Research Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.3 Research Findings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.4 Complicating Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.4.1 Model Conceptualization . . . . . . . . . . . . . . . . . . . . . . . . 36
2.4.2 Myriad of Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . 37
ix
x Contents
1.1 Introduction
5 cell 3 7
3
5
Workstation 2
1. Every job operation is performed on one and only one machine at any given
point in time;
2. The processing times of the operations are machine-dependent, and the
machines are independent from each other;
3. Once started, each operation must be performed to completion without
interruption;
4. Setup times are of machines are included in the job processing time of the job;
and
5. The transportation time of jobs between machines are negligible and are
included in the processing times.
Table 1.2 illustrates a job shop problem with 3 machines to process 3 jobs upon
which 8 operations are to be performed. For instance, operation 11, 12, and 13
represent the first, second, and third operations of job 1, respectively. Machines
capable of performing each operation are as shown. The objective is twofold: (i) to
assign each operation to an appropriate machine, which is a routing problem, and
(ii) to sequence the operations on specific machines, which is a sequencing prob-
lem, so that make-span and the total working time of machines (total workload) are
minimized. This becomes a complex multi-criteria optimization problem that
demands significant computational resources.
Figure 1.3 represents a typical solution to the problem, where operations 11, 13,
and 31 are performed on machine m1, operations 21, 23, and 12 are performed on
m2, while 22 and 32 are done on m3. A group representation scheme of the solution
is shown in (b).
Figure 1.3 represents a typical solution to the problem, where operations 11, 13,
and 31 are performed on machine m1, operations 21, 23, and 12 are performed on
m2, while 22 and 32 are done on m3. A group representation scheme of the solution
is shown in (b).
Table 1.2 A job shop problem with 3 machines and 3 jobs with 8 operations
Jobs J1 J2 J3
Operations 11 12 13 21 22 23 31 32
Machines m1, m2, m3 m1, m2 m1, m2, m3 m1, m2, m3 m1, m3 m1, m2, m3 m1, m2, m3 m2, m3
10 1 Exploring Grouping Problems in Industry
m2 m3
13 6
40 20 21
18 12
1 7
21 18
0
w3
1
0
In bin packing, objects of different volumes and shapes must be packed into a finite
number of bins, where each bin has a specific volume. Oftentimes, the goal is to
ensure that the wasted space or number of bins used are minimized as much as
possible (Allen et al. 2011; Pillay 2012). According to computational complexity
theory, the bin packingproblem is a NP-complete problem that is computationally
expensive and highly combinatorial, when formulated as a decision problem (Pillay
2012). Figure 1.6 presents an illustrative example of the bin packing problem and
its group structure. In part (a), the problem shows three bins, that is, b1, b2, and b3,
which are packed with groups of objects (1, 4, 8), (5, 6), and (2, 3, 7), respectively.
This information can be presented conveniently in a group structure as indicated in
part (b).
1 7
3 4
5
Worker w2
2006; Chen 2007; Tarokh et al. 2011). It is desirable to develop heuristic algorithms
that take advantage of the grouping structure of the problem.
Modular design is a design approach that divides a system into modules that can be
independently created for assembling a variety of different systems. There are three
basic categories of modular design, namely function-based modular design, man-
ufacturing design, and assembly-based modular design (Tseng et al. 2008).
Modular design is important as manufacturers need to cope with multiple variations
of product specifications and modules in a customized environment (Kreng and Lee
2003). Thus, well-developed modular designsystems will help with the production
and control of mass customization. Modular design also focuses on environmental
aspects based on grouping techniques.
In the assembly-based modular design method, products are generally described
by liaison graph (Tseng et al. 2008). Figure 1.8 shows an example of a Parker Pen
assembly with 6 components to be grouped into 3 modules. There are three
important stages in modular design:
1. Determination of liaison intensity (LI) of components;
2. Grouping or clustering of components using a grouping method; and
3. Evaluation of the clustering of grouping result.
The goal is to maximize the liaison intensity within each module and to mini-
mize the liaison intensity between modules. However, as the number of components
and modules increase, the number of possible combinations increases
1.Button
2 Group G2
components at one goal. This means that specific groups of components or sub-
systems have to be cautiously defined so that each group can undergo preventive
maintenance within a defined time window. The overall aim is to minimize
maintenance costs (setup, preparation costs) while maximizing the reliability of the
systems. This grouping problem is twofold (Dekker et al. 1997):
1. Fixed group models, where all components are always jointly maintained as a
group; and
2. Optimized-groups models, where several groups are optimally generated, either
directly or indirectly.
The major challenge in solving the group maintenance problem is its compu-
tational complexity due to exponential growth of the number of variables as the
number of components or subsystems increase. Efficient and robust metaheuristic
methods are a potential option.
"Both General Sumner and you sent me word to hold the line at
all hazards, and that night we dug a line of intrenchments
across our front, using the captured Spaniards' intrenching
tools. We had nothing to eat except what we captured from the
Spaniards; but their dinners had fortunately been cooked, and
we ate them with relish, having been fighting all day. We had
no blankets and coats, and lay by the trenches all night. The
Spaniards attacked us once in the night, and at dawn they
opened a heavy artillery and rifle fire. Very great assistance
was rendered us by Lieutenant Parker's Gatling battery at
critical moments; he fought his guns at the extreme front of
the firing line in a way that repeatedly called forth the
cheers of my men. One of the Spanish batteries which was used
against us was directly in front of the hospital so that the
red cross flag flew over the battery, saving it from our fire
for a considerable period. The Spanish Mauser bullets made
clean wounds; but they also used a copper-jacketed or
brass-jacketed bullet which exploded, making very bad wounds
indeed.
"Since then we have continued to hold the ground; the food has
been short; and until today [July 4] we could not get our
blankets, coats, or shelter tents, while the men lay all day
under the fire from the Spanish batteries, intrenchments, and
guerrillas in trees, and worked all night in the trenches,
never even taking off their shoes. But they are in excellent
spirits, and ready and anxious to carry out any orders they
receive. At the end of the first day the eight troops were
commanded, two by captains, three by first lieutenants, two by
second lieutenants, and one by the sergeant whom you made
acting lieutenant. We went into the fight about 490 strong; 86
were killed or wounded, and there are about half a dozen
missing. The great heat prostrated nearly 40 men, some of them
among the best in the regiment."
"The orders under which the battle of San Juan was fought were
given by Adjutant-General McClernand to General Kent, commanding
the Infantry Division—consisting, in addition to the
organizations already mentioned (Wikoff's and Pearson's
brigades), of the First Brigade, including the Sixth and
Sixteenth United States Infantry and the Seventy-first New
York, under General Hawkins—at about nine o'clock in the
morning. There is no question fortunately as to the exact
wording of the orders. A little green knoll to the left of the
Santiago road and half a mile short of the San Juan Heights was
pointed out as the point which was to be the extreme limit of
the forward movement of the Infantry Division. Once there,
further orders would be given. The orders under which General
Sumner advanced from El Pozo would appear to have been more
specific, and certainly more clear than the orders which
General Kent received for the Infantry Division a few minutes
later. At the same time, it is true that these orders were
also based upon a complete misconception of the situation and
a total ignorance of the Spanish position and the lay of the
country beyond El Pozo. General Sumner's orders were to
advance along that branch of the Aguadores Creek which runs
parallel with the Santiago road from El Pozo, until it joins
the main stream of the Aguadores at the angle subsequently
known as the 'bloody angle,' where the creek makes a sharp
turn to the left, and then runs a general southerly course
toward the town of Aguadores and the sea. This creek General
Sumner was instructed to hold until the result of Lawton's
attack upon Caney became known, and he received further
orders. Once the creek was reached, Sumner, under the most
unfavorable circumstances of a heavy fire, and the thick and
pathless jungles which his men had to penetrate, deployed his
whole division, and then sent back word to McClernand, the
adjutant-general of the corps, acquainting him with the actual
conditions by which he was confronted, and asking whether his
orders contemplated an attack upon the enemy's intrenched
position, setting forth at the same time the utter
impossibility of keeping his men inactive for a long time
under such a heavy fire as was being poured in upon them.
{607}
S. Bonsal,
The Fight for Santiago,
chapter 6 (New York: Doubleday, Page & Company).
"When the first-aid bandages were applied, the wounded man and
those who helped him were, as a rule, under fire, which made
any but the most summary methods of wound-dressing quite
impossible. Fortunately these bandages, so simple and
practical, lent themselves excellently well to this procedure.
The first thing the soldiers or the hospital attendants would
do when they came upon a wounded man was, in the case of a
wound in the body, to tear off his shirt, or in the case of a
wound in the leg, tear off his trousers, and then wrap around
the wound the first-aid bandage. The wound-dressers were
generally in such haste, and the wounded men usually so
helpless to assist in any way, and their shirts and trousers
so rotten from the drenching rains in which they had been worn
without change day or night, that the taking off of the clothing
was literally what I call it—tearing, and the garment came off
so rent as to be quite useless for further wear. Consequently
the soldiers were carried half-naked, or, if they had been
wounded in both the body and the lower limbs, as was so
frequently the case, entirely naked, to the army wagons and so
down to the hospital, where there was not a scrap of clothing
or bedding forthcoming to cover them with. These who were
stripped in this way during the daytime were baked and
blistered by the fierce sunlight, only to shiver with the
penetrating cold and dampness after the rain had ceased to
fall and when the chill night came on.
S. Bonsai,
The Fight for Santiago,
chapter 8 (New York: Doubleday, Page & Company).
The troops which had carried San Juan Hill were intrenched,
that night, in the positions they had gained, and those which
had taken El Caney were brought into connection with them,
Lawton's division on their right and Bates's brigade on the
left. The battle was renewed by the Spaniards soon after
daylight on the morning of the 2d, and raged with more or less
fury throughout the day. That evening, about 10 o'clock, a
fierce attempt was made to break through the American lines,
but without success. Again, on the morning of the third, the
Spaniards reopened battle, but with less vigor than before.
General Shafter then sent the following letter to General
Toral, the Spanish commander: "I shall be obliged, unless you
surrender, to shell Santiago de Cuba. Please inform the
citizens of foreign countries, and all women and children,
that they should leave the city before 10 o'clock to-morrow
morning." In reply, General Toral wrote; "It is my duty to say
to you that this city will not surrender, and that I will
inform the foreign consuls and inhabitants of the contents of
your message." Several of the foreign consuls at Santiago then
came into the American lines and persuaded General Shafter to
delay the shelling of the town until noon of the 5th, provided
that the Spanish forces made no demonstration meantime against
his own. This established a truce which was renewed, in a
series of negotiations until the 10th. "I was of the opinion,"
reported General Shafter, "that the Spaniards would surrender
if given a little time, and I thought this result would be
hastened if the men of their army could be made to understand
they would be well treated as prisoners of war. Acting upon
this presumption I determined to offer to return all the
wounded Spanish officers at El Caney who were able to bear
transportation, and who were willing to give their paroles not
to serve against the forces of the United States until
regularly exchanged. This offer was made and accepted. These
officers, as well as several of the wounded Spanish privates,
27 in all, were sent to their lines under the escort of some
of our mounted cavalry. Our troops were received with honors,
and I have every reason to believe the return of the Spanish
prisoners produced a good impression on their comrades. The
cessation of firing about noon on the 3d practically
terminated the battle of Santiago." General Shafter goes on to
say that when the battle was fiercest, on July 1st, he probably
had no more than 12,000 men on the firing line, not counting a
few Cubans who assisted in the attack on El Caney, and who
fought with valor. They were confronted by about equal numbers
of the enemy, in strong and intrenched positions. "Our losses
in these battles were 22 officers and 208 men killed, and 81
officers and 1,203 men wounded; missing 79. The missing, with
few exceptions, reported later." Up to this time, General
Shafter had been unable to complete the investment of the town
with his own men, and had depended upon General Garcia with
his Cubans, placed on the extreme right of the American lines,
to watch for and intercept reinforcements. They failed to do
so, and 2,800 Spaniards, under General Escario, entered the
city on the night of the 2d. The American commander now
extended his own lines as rapidly as possible and completed
the investment of the town.
Report of Inspector-General
(Annual Reports of the War Department, 1898,
volume 1, part 2, page 596).
"The enemy's vessels came out of the harbor between 9.35 and
10 a. m., the head of the column appearing around Cay Smith at
9.31 and emerging from the channel five or six minutes later.
The positions of the vessels of my command off Santiago at
that moment were as follows: The flagship 'New York' was 4
miles east of her blockading station and about 7 miles from
the harbor entrance. She had started for Siboney, where I
intended to land, accompanied by several of my staff, and go
to the front to consult with General Shafter.
{609}
A discussion of the situation and a more definite
understanding between us of the operations proposed had been
rendered necessary by the unexpectedly strong resistance of
the Spanish garrison of Santiago. I had sent my chief of staff
on shore the day before to arrange an interview with General
Shafter, who had been suffering from heat prostration. I made
arrangements to go to his headquarters, and my flagship was in
the position mentioned above when the Spanish squadron
appeared in the channel. The remaining vessels were in or near
their usual blockading positions, distributed in a semicircle
about the harbor entrance, counting from the eastward to the
westward, in the following order: The 'Indiana' about a mile
and a half from shore, the 'Oregon'—the 'New York's' place
being between these two—the 'Iowa,' 'Texas,' and 'Brooklyn,'
the latter two miles from the shore west of Santiago. The
distance of the vessels from the harbor entrance was from 2½
to 4 miles, the latter being the limit of day blockading
distance. The length of the arc formed by the ships was about
8 miles. The 'Massachusetts' had left at 4 a. m., for
Guantanamo for coal. Her station was between the 'Iowa' and
'Texas.' The auxiliaries 'Gloucester' and 'Vixen' lay close to
the land and nearer the harbor entrance than the large
vessels, the 'Gloucester' to the eastward and the 'Vixen' to
the westward. The torpedo boat 'Ericsson' was in company with
the flagship and remained with her during the chase until
ordered to discontinue, when she rendered very efficient
service in rescuing prisoners from the burning' Vizcaya.' …
{610}
"I regard this complete and important victory over the Spanish
forces as the successful finish of several weeks of arduous
and close blockade, so stringent and effective during the
night that the enemy was deterred from making the attempt to
escape at night, and deliberately elected to make the attempt
in daylight. The object of the blockade of Cervera's squadron
was fully accomplished, and each individual bore well his part
in it—the commodore in command on the second division, the
captains of ships, their officers, and men. The fire of the
battle ships was powerful and destructive, and the resistance
of the Spanish squadron was, in great part, broken almost
before they had got beyond the range of their own forts. …
Several of the [American] ships were struck—the 'Brooklyn'
more often than the others—but very slight material injury was
done, the greatest being aboard the 'Iowa.' Our loss was 1 man
killed and 1 wounded, both on the 'Brooklyn.' It is difficult to
explain this immunity from loss of life or injury to ships in
a combat with modern vessels of the best type, but Spanish
gunnery is poor at the best, and the superior weight and
accuracy of our fire speedily drove the men from their guns
and silenced their fire. This is borne out by the statements
of prisoners and by observation."
{611}
"While one of our boats was still ashore, seeing heavy clouds
of smoke behind the next point the ship was moved in that
direction, the men being at quarters and everything in
readiness for further action. On rounding the point two
men-of-war were found on the beach burning fiercely aft, the
majority of the crew being crowded on the forecastle and
unable apparently to reach land, only 200 yards away. Our
boats, under Lieutenant Norman and Ensign Edson, put off to
the nearer vessel, which proved to be the flagship 'Infanta
Maria Teresa,' and rescued all on board by landing them on the
beach through the surf. Lieutenant Norman formally received
the surrender of the commander in chief and all his officers
and men present, and as soon as all hands had been transferred
ashore, brought on board this ship all the higher officers,
including the admiral. Lieutenant Wood meanwhile rescued the
remaining survivors on board the 'Oquendo,' the second of the
burning vessels. The Spanish officers not feeling that the
prisoners on shore were secure from attack by Cuban partisans,
by your orders I directed Lieutenant Norman to land with a
small force, establish a camp on shore, and hoist the United
States flag over it. He took with him all the rations that
could be spared from the stores aboard."
"On July 8 General Toral offered to march out of the city with
arms and baggage, provided he would not be molested before
reaching Holguin, and to surrender to the American forces the
territory then occupied by him. I replied that while I would
submit his proposition to my home Government, I did not think
it would be accepted. In the meanwhile arrangements were made
with Admiral Sampson that when the army again engaged the
enemy the navy would assist by shelling the city from ships
stationed off Aguadores, dropping a shell every few minutes.
On July 10 the 1st Illinois and the 1st District of Columbia
arrived, and were placed on the line to the right of the
cavalry division. This enabled me to push Lawton further to
the right and to practically command the Cobra road. On the
afternoon of the date last mentioned the truce was broken off
at 4 p. m., and I determined to open with four batteries of
artillery, and went forward in person to the trenches to give
the necessary orders; but the enemy anticipated us by opening
fire with his artillery a few minutes after the hour stated.
His batteries were apparently silenced before night, while
ours continued playing upon his trenches until dark. During
this firing the navy fired from Aguadores, most of the shells
falling in the city. There was also some small-arms firing. On
this afternoon and the next morning we lost Captain Charles W.
Rowell, 2d Infantry, and 1 man killed, and Lieutenant Lutz, 2d
Infantry, and 10 men wounded. On the morning of July 11 the
bombardment by the Navy and my field guns was renewed and
continued until nearly noon, and on the same day I reported to
the Adjutant-General of the Army that the right of Ludlow's
brigade of Lawton's division rested on the bay. Thus our hold