~ eonstructlon Mana ement SPPU 9-2

,, 5teven Tanimoto - In the middle section of artificial neuron is sum

•'Computational techniques for performing tasks that function that sums all weighted inputs and biu.
:,. CHAPTER apparently require intelligence when performed by - At the exit of artificial neuron the sum of previously
humans." weighted inputs and biaa is passing through activation
Introduction to Artificial Intelligence.
Technique
., David Pama•
"Artificial intelligence is to artificial flowers as natural
- - - - -. _.,. _____
function that is also called tranJfer function.

intelligence is to natural flowers."

,.- Luger .
The branch of computer science that is concerned with
automation of intelligent behaviour.
,.- Rich
Syllabua
_ "Al is the study of how to malce computers do things
Basic terminologies and applications In civil engineering
which, at the moment, people do better."
(a) Arllflclal neurai network (b) Fuzz! !ogle (c) Genetic algorithm
9.2 Artlflclal Neural Network
~ (SPPU • May 12, May 16)

complex problems to the levels of experts by means of s- Introduction

9.1 Artificial Intelligence (Al)
imitale experts. Artificial neural network are the mathematical
~ (SPPU - May 12, May 1', May 18) _ inventions motivated by observations made in study of
All in all, artificial intelligence has a broad application
Artificial intelligence, a comprdlemive discipline, was prospects in the practice of civil engineering. The biological system, though loosely founded on the actual Flg. 9.2.1 : Worldn1 prtndple or an artlfldaJ neuron
developed bued on the intaaction of seven) kinds of scientific goal of AI is to understand the principles that biology.
diJciplinea, such as computer science, cybemctics, Although the artificial neuron's working principles ~d
make intelligent behavior possible · in natural or An artificial neural network can be defined as.mapping - set of rulea are simple, the full potential and calculation
information theory, psychology, linguistics, and artificial systems. ThiJ is done by 1) the analysis of an input space to output space.
neuropfiysiology. power of these modela come to life when _it
oatun1 and artificial agents. 2) FO!lDulating and teating 1bis concept is analogous to that of mathematical interconnect into artificial neural networks shown 10
- Artificial Intelligence (AI) is a branch of Science which hypotheaea about what it takes to construct intelligent function. Fig. 9.2.1.
deals with helping macbiaes finda 10luti0111 to complex agents; and 3) Designing, building, and experimenting - The purpose of neural network is to map an input into _ These artificial neural networks use simple fact that
problmm in a more-human -lite fubioo. Thia geoenl.ly with computatioruil systems that perform tasks desired output. complexity can g,:ow out of merely few baaic and
involves bom>wing chanicteristica from human commooly viewed as mquiring int.elligence.
intelligence, and applying them u algorithms in a While patterned after the interconnections between simple rules.
- Io structural mechanics and donstruction materials neurons found in biological system, artificial neural
computer friendly way.
contexts, recent experiments have reported that fuzzy oetw~rk are no more related to real neurons than
- A more or less flexible or efficient approach can be
logic (FL), artificial neural networb (ANNs), genetic feathers are related to modem airplanes.
taJcen depending OD the mjuiremeots established, algorithm (OA), and fuzzy genetic (PG) may offer a
which influences how artificial tbe intelligent bebavior Biological systems, neurons and feathers, ~rv~ a
appears.
promising alternative. useful purpose,but the implementation of pnnc1plea
- They are known as artificial int.elligence (AI). involved has resulted in man-made inventions that bear input
- In the field of civil engineering, many problems, loyer
especially in engineering deaign, conatructioo - In civil engineering, AI methods have been extensively little resemblance to biological systems that spawned
management, and program decision-making, were
influenced by many uncertainties which could be
used in the fielda of civil engineering applications such
88
COOStructioo management, building materials, -
the creative Process•
Basic building block of every artificial neural netw~rk
e st,g1eneuran

hy~c, oplimiution, geotechoical and transportation is artificial neuron, that is; a simple mathematical fi1g. 9.2.l: Example olllmple utlftdal neural network
solved not only in need of mathematics, physics, and
mcclianics calculations but also depend on the
C08111Caing.
model (function). _ In order to have the benefits of mathematical
experience of pnctitiooen. 9 Such a model has three simple sets of rules: complexity that can be attained through interconnection
•1.1 Some deffnltlona of Al . d activation. At the
This .knowledge and experience are illogically multiplication, summation an . led of individual artificial newoos and not just making
• Wlnaton
incomplete and impreciae, and they canoot be bandied entrance of artificial neuron the inputs are weigh . system complex and wunaoageablc, usually artificial
by traditional procedures. However, artificial - "AI is the · what means that every input value 11 · ul · lied with neurons cannot be iotercoMected randomly.
thin IIUdy of Ideas which enable computers to do m bp
intelligence has its own superiority. It can solve gs which make people seem intelligent." indi~idual weight.
 ~ ConatnJction Mana ement SPPU)
Introduction to Artlflclal lntell ence T8Chn
lier M eonetructlon Manapement (SPPU)
- There are severa) "atandardir.ed" topographies of
artificial neural networb. These ~fined
9-3
Hidden layers are layers betWeen the input and oUIJ>ut
layers. The neW'Ons in ~h layer are uauauy full~
connected to the neurons m another layer. The three.
-- Toe input units are directly COllnected with the output
unit anatid they · are genenJly uaed for pattan
9-4 lnlloduction to Artlllclal lntellgenee Techn~ue r
.., Wortllng

A random unit ii ch01en. If any of itJ neighbors are

ciualfic on.
topographies are helpful in solving problem eaaily, layer feed forward network is the moat popular illd active, the unit computea the ,um of the welghtJ on the
speedily and more efficiently. CODncctiODI to those active neighbors. If the ■um 11 •
mostly used.
- These several artificial neural network topographies are _ The· dala proceaslng can extend over multiple Oayen poeitive result, then the unit becomes active, otherwise it
,..._-.._....,,.. Output Layer
appropriate for solving diffecent types of problems. becomes inactive.
of) unil5, but no feedback connections are Pltaenl
- After defining the type of given problem lhc fir!t step is Recurrent networks comprise · feedback COnnectiona,
to decide for topology of artificial neural network and Input Layer
- Feature
Contrmy to feed-forward networlcs, the dyn811!Jcl! Fault-tolennce: If a new procewng element fails
then fine-tune it along with itJ panmcter. Uke properties of lhc network are important. completely, the network will ■till function properly.
biological neural netwodcl which can learn their
In some cases, the activation values of the uniti
behavior /responses on the buis of inputJ that they get - 9.4 Features of Artlflclal Neural Network
undergo a relaxation process such that the network Will Fis- 9.3.2 : Slnp-layer feed forward network
from their environment the artificial neural networks
can do the wne. evolve to a stable state in which these activationa do • A~Na have lollowlng fNturN
. . (b) Multilayer network
not change anymore. '
- There arc three major learning paradigms: supervised · A multilayer network hu more layen of nodes between - Their ability to repreaent non-linear relations make,
learning, unaupcrvised learning and mnforcemcnt In other applications, the changes of the activation
the input unita and output unilJ. Typically there ls a them well lllited for non-linear modelling in control
learning. values of the output neurons are significan·1, such that
layer of weightJ between 2 adjacent levels of unitJ. ■yatems.
the dynamical behavior constitutes the output of the
- Bucd on the problem which is. to solve, learning Adaptalion and learning in WJCel'lain system through
Multilayer nets can solve more complicated than
paradigm can be chosen like as artificial neuron nelWOIL XI
single-layer nets, but ita training is more difficull off line and on line weight adaptation.
network topography.
- Parallel proceuing architecture allows fut proceaaing
- Although learning paradigms are dissimilar in their
for large-scale dynamic system.
. principles they all have one thing in common; on the
baai1 of "learning data" and "learning rules" (chosen - Neural network can handle large number of lnputJ and
cost function) artificial oeuraJ network is bying to I can have many oulp\ltl.
achieve proper output response in accordance to input ANNs can store knowledge in a distributed fuhion and
Fig. 11.3.I: Architecture or artUldal neuron
lignals.
Input Layer J- consequently have a high fault tolerance.
- Artificial neural netwodcs are working in many - Based on the no of layers the network structure is
different arcu such u proceu control, chemistty, broadly classified into two categories depends on the
9.5 Features of Blologlcal Neural
gaming, radar systems, automotive industry, space
Fis- 11.3.3 : Multilayer feed forward network fully Networks
connection establishment in the network. connected
indultry, utronomy, genetics, banking, fraud detection,
Nllwortt Structu,.
1 Some allrletive featurea of the biological neural
etc. and solving of problems like function ~ 9.3.2 Recurrent Network
networlc that make it superior to even the most sophisticated
approximation, regrmon analyli1, time series
- Here each node is a ~ing element or unit; it may artificial in1Clligcnce computer system for pattern
prediction, clauificalion, pattern recognition, decision 1. Feed-folward netwo,k be anyone of. the 2 states (Black-Active, White- recognition tub are the following :
making, data procaaing, filtering, clustering, etc.
(a) Single layer nelwolk
Inactive). Units are connccied to each other with
9.3 Neural Network Architectures weighted symmetric connection.
(b) Mijd layer netwof1c - A positive weighted connection indicates that 2 unil5
~ (SPPU • Mey·12, Mey 14, Mey 11)
tend to activate each other. A negative connection ( a ) ~ and lauN toleranOII
- A common ANN architecuR is determi.ncd by three 2. Recurrent netwolic
allows an active unit to deactivate from a neighbouring
distinguiahing characteristics: connection types, (b)Fllxl,llty
connection schemes and layer configunti0111. ~ C!l.l : Feed-forward network unit.

.. 9.3.1 Feed-Forward Network (c) ~ to dNI wtlh a variety ct

- It 00llliJII of three typea of nairon Iayen: input, data alluatlona
hidden, and output layen. In feed-forward nctworb, . . (a) Single layer network
the signal flow is from input to output unitJ, strictly in a _ (d) CollectlV9 computallon
feed-forward di=tion. It ~ only one layer of connection weighta. Toe input
Input Layer
umt will receive signals from the outaide world and the Fis, CU I Featunl of blolopcal NN
- Input layer is the layer that receives .input lignalJ from
response to• signal can be read from the output uniL
the environment Output layer is the layer that emitJ Fla- 11.3.4 : Recurrent network wltb DO ,elf.feedback
signal, to the environment. loopl
 ~ ConatJuction Manap:m-rt (SPPIJ) 9-5
I n ~ 10 Artificial Intelligence Techn~ '
lf!: constrUCflon Management (SPPU) 9-6 Introduction to Artlflcial tntellgenoe Techn~u• r
,, l)lfldrtte ;,,,...s interconnections
➔ (a) Robmtma and fault tolerance Bach neuron baa fine. hair like tubular •1ructurei 9.. 72 intcn:onoectioos with appropriate weights, the
The decay of nerve ceU. does not seem to affect the - In an artificial neural network several J)IOcesaing units sctivation 11ates detamine the abort term memoiy
(exteDJions) around iL function of the network.
~ormance significantly. are interconnected according to some topology to
They t,nnch out into tiee around the cell body. 'Ibey accomplish a pattern recognition task.
➔ (b) lileldblllty - Generally, given an external input, the activation
accept inCOming signals.
,The network automatically adjlllll to a new 11terefore the inputs to a procesaing unit may come dynamics is followed to recall a pattern stored in a
environment without using any pq,rogrammed rr Axon from the outputa of other processing units, and/or from networlc. In Older to store a pattern in a network. it ii
instructiona. . long thin tubular structure which works
- It IS I • '
u~-
""l external source&. ·neceswy to adjuat the weights of the connections in
➔ (c) AbWty to deal with a variety of data altuatlom transmission line. 11te output of each unit may be given to several units the network.
The network can deal with information that is fuzzy, ..- Syflll)lt including itself. The amount of the output of one unit
- The set of all weights on all connections in a network
probabillJtic, noisy and inconsistenL Neurons are connected to one another in complex · received by another unit depends on the slicngth of the
fonn a weight vectm. The set of all posaible weight
➔ (d) Collectne compidalloa' · connection between the unita, and it ia reflected in the
spatial arrsngemenL vecton define.the weight space.
The network perf~ routinely .many operations in When axon reaches ita final destination it branchca weight value associated with t h ~ n g link.
When the weight, are changing, then the synaptic
parallel and a11o a fven
task in a distributed manner. - again called u tenninal arborization. At the end of If there are N unita in a given , en at any iii.slant
dynamics of the networlc determines the weight vector
axon are highly complex and specializ.ed structure& of time each unit will have a unique tivation value
9.6 Biological Neural Networks
called synapees. Connection between two neurons takes and a unique output value. as a function of lime. Synaptic dynamics ia followed to
+ (SPPU. Illy 12, Illy 11) place at lheae synapeea. The set of the N activation values of the network adjust the weight, in order to store the given patterru in

- The featurea of lhe biological neural netwoat are defineS the activation state of the network at that the network.
Dendrites receive the input through the synapses of
attributed to ill IIIUcnn and function. instanL Likewise, the set of the N output values of the _ The process of adjusting the weighl8 is refemd to u
Olher neurons. The soma proceases these incoming
netWork definea the output state of the networlc at that learning. Once the teaming process Is completed, the
- The fundamental unit of the network ia called a IICUIOll signa)J over lime and converts that prooesaed value into
or • nerve cell. Fig. 9.6.lshows a schematic of lhe an output. which i1 sent out to other neurons through instant.'° final set of weight values corresponds to the long term
llrUclure of a neuron. lhe axon and the synapses. - Depending on the discrete or continuous nature of the memory function of the network.

FIOITI-' 9.7 Artlflclll Neural Networks:

activation and output values, the state of the network
can be described by a discrete or continuous point in an -
The procedure to incrementally update each of the
weights ia called a learning law or learning algorithm.
Tennlnology N-dimensional space.
9.7.4 Update
9.7.1 Procealng Unit 9.7.3 • Operations
In implementation, there are seven! options available
In operation, each unit of an ANN receives inputs from
We can consider an artificial neural network (ANN) as for both activation and synaptic dynamics.
other connected·units and/or from an ~xtcmal source. A
a highly simplified model of the structure of the weighted sum of the inputs is computed at a given - In particular, the updating of the output states of all the
biological neural network. An ANN consist, of instant of time. unil8 could be performed synchronously.
intaconnected processing units. The general model of a - The activation value determines the actual output from _ In this case, the activation values of all the unita are
NEIRlN2 proceuing unit consist, of a summing pan followed by the output function uni~ i.e., the output state of the unit computed at the same time, .assuming a given output
F1a. ll.6.1: Sc:be-tlc dlaenm ota typlcal nmroa or an output part. The output values and. other external inputs in tum state throughouL From the activation values, the new
oenec:eD determine the activation and output states of the other output state of the networlc ia derived.
- The summing part receives N input values, weights
- Within hlllDIIIJ there are many variali0111 •on basic units. In an asynchronous update, on the other hand, each unit
each value, and computes a weighted sum. The
type of neuron, yet, all biological IICUl'OIII have the
- Activation dynamics determines the activation values is updated sequentially. talring the cwrent output state
weighted sum is called the activation value. The output of all the units, i.e., the activation state of the network
aame four buic component,. They e identified by of the networlc into account each time. For each unit,
part produce& a signal from the activation value. as a function of time. The activation dynamics also
their biological name, - cell body (soma), dendrit.ea, the output state can be determined from the activation
axon, and.1ynapeea. - The 1ign of the weight for each input determinea determines the dynamics of the output state of the value either deterministically or stochastically.
• Cell body (Soma) whether the _input is excilaloiy (positive weight) or network. In practice, the activation dynamics, including the
inhibitory (negative w~ight). The set of all activation states defines the activation - update, is much more complex In a biological neural
The body of neuron cell CODtaim the nucleus and
state space of the networlc. The set of all output statcs networlc than the simple models mentioned above. The
carries out biochemical tr&Dlformalion necessary to the - The inputs could be dilcrete or continuous data values, defines the output state space of the network. ANN models along with tho equations governing the
life of neurons.
and likewise the outputs also could be discrete or - Activation dynamics determines the trajectory of the activation and synaptic dynamics are designed
.contin'!OU'· The input and output could also be path of the states in the state space of the network. ~ according to the pattern recognition task to be handled.
deterministic or stochutic or fuzzy. a · ven network, defined b the units and their
w~ Mana=
9.8
ISPPU)
Charactert1t1ca ot Neural Networb
9-7
I ~ to Artlllclal lnteUlgence T ~

selJ allow for quick proceuing of infOl'lnatiori by

~ construction Management (SPPU)
► Soft computing contaim fuzzy logic, DeUral networks,
9-8

_
Introduction to Ar1lflclal lntellgence Techn~ue r
As an outcome; in many CIICI it ia capable of aolving
- ::.Uon of vaguely similar patterns while PIOVidin
- probabilistic ~ g , and genetic algorithms. the problem in appropriate way. Even milling input
mean, to deal scientifically with subjcctiVity g
- The oewaJ nctworb exhibit mapping capabilities, lhat die . h 'all •1 Combination of techruques from all theac areas are information has even now turned out to be aatiafactory
i1, they can map input pa11em1 to their mociatcd tenitOJY lhat traditional science u eaacnli y iSllortd_
UJCd to design an intelligence system. in knowledge-based syatema.
output pattern,, In other words, fuzzy. l°".c is concerned ~th Plllting Neural networks · provide algorithms for learning, - Puzzy logic pcmrill to lower complexity by allowing
- The neural networb 1ecn by examplea. Th111, neural away from logic that 1s cnsp or Boolean (binary; 0 Or the use of imperfect data in SC11Jible way. It can be
classification, and optimization, whereas fuzzy logic
network architectura can be 'lraincd' wilh known I). implemented in hardware, aoftware, or a combination
dea)s with issues such aa fonning impreasiona and
examplca of a problem before Ibey are tested for their of both. In other words, fuzzy logic approach to
Thia melhod has an advantage over Boolean logic io reasoning on a semantic or lingui1tic level.
'inference• capability on unknown U1J1aDCea of die - problema' control mimics how a pmon would inlJce
problem. They can, therefore, identify new objecll 11111 it mi.mica complex bwnan reasoning so aa to lltive Probabilistic reasoning deals with uncertainty. decisions, only much faster.
previously Ulllrained. 11 realistic conclusiona about the imprecise and often Although there are significant areas of overlap between The fuzzy logic analysis and control methods shown in I
- The neural nctworb poucu the capability to
fuzzy niture of reality.
neural networks, fuzzy logic, and probabilistic Fig. 9.10.1 can be described u : •
gcocnliz.e, Th111, Ibey can jedict new OUIComea from - The past few yem have observed a rapid &rowth in the reasoning. In general they are complementary rather • Receiving one or large number of measurements
pultmida. number and variation in applicationa of fuzzy IOgic than competitive.
1 or other asseumcnt of conditions existing in some
- The neural nctworb are robuat syslems and me fault (FL). Fu1.zy logic methods have been uaed in image- Recently, several intelligent systems called neuro-fuzzy ayatem that will be analysed or controlled.
tolmnL They can, thcref<n, recall fulJ pa11an1 from Ulldtntanding applicationa such u detection of cdg~
incomplete, partial or noisy )llllans. systemB have been used. There are many methods to Processing all received inputs according to human
feature extraction, clwilication, and clustering. combine neural networks and fuzzy logic techniques based, fuzzy "if-then" rules, which can be
- The neural networb can )IIOCCl8 infOllllllion in
- Fu1.zy logic p01ca die capability to mimic the hllll]an however, it is essential to understand basic ideas in the expreased in simple language words, and
paRIJcl, at high lpeed, and in a dillribured IIIIIWr,
mind to effectively employ means of l'eaSoning that art design of fuzzy logic techniques. combined with traditional IIOll•fuzzy procasing.
9.9 Fuzzy Logic approximate rather than exact. In traditional hard 1

Characteristics of Fuzzy Logic Averaging and. weighting the results from all the
computing, decisions or actions are based on Jncisioo, 9.9.1
. . (SPPU • May 12)
individual rules into one aingle output decision or
certainty, and vigor.
• Introduction The casential characteristics of fuzzy logic are related signal which decides what to do or tella a
- Precision and certainty carry a coat In soft computing, to the following: controlled system what to do. The result output
- Fuzzy Logic WU initialed in 1965 by Lolfi A. 2.adeh, tolmnce and impreaaion are explored in decision In fuzzy logic, exact reasoning is viewed as a limiting signal is a pm:isc defuzzllied value.
profcuor for computer ICiencc • the Univmity of making.
California in Bakeley. cue of approximate reasoning. The following is Fuzzy Logic Control/Analysis Method
- The exploration of the tolerance for imprecision and In fuzzy logic, everything is a matter of degree. diagram.
- Basically, Fuzzy l.ogjc is a multivalued logic, which uncertainty IIDdediea the reIDarkable hwnan ability to
allow, intcmiediatc valuea to be defined between In fuzzy logic, knowledge is inle!pretcd a collection of
uodentand dilto!ted speech, decipher sloppy elastic or, equivalenlly, fuzzy constraint on a collection
conventional ellaluati0111 like true/false, vea/no
handwriting. comprdiend DIWICca Of natural
high/low, etc. Notions like radiec tall or vay fl8;can ~ of variables.
language, SIIDIDlariu text, and RCOgniz.e and clllsify
formulated DWhcrn•tically 111d !lroceaed by imagea. Inference· is viewed as a process of propagation of
80 18
cotnputm, to apply • IDOl'e human like way of elastic constraints.
thinking in the PIOgrlmming of COIDpUlcn, - Wilh Fuuy logic, mapping rules can be specifying in
Any logical system can be fuzzified.
- Puzzy logic is a comp1cx 11111bema1icaJ method lbat IClma of Words rather than nwnbers. Computing wilh
the w<Xds cxplorea Ullpreciaion and tolerance. In Fuzzy Fl&,9.lo.t
~owa aolving difficult limuJatcd problems with many 9.10 Fuzzy Logic System
1Dpull and output Variablca. logic, another basic concept is the fuzzy if-then rule. _ With the purpose of operating fuzzy logic needs to
- :'11hough rule-baaed systems have a long hiatOJy of use Today control systems are usually described by be represented by numbcn or descriptions. For
- Fuzzy logic iJ capable to give l'elUIII in the form of mathematical models that follow the laws of physics,
in artificial intelligence, but machinery for dealing with example, speed can be represented by value 5 mis
~ ~ o n for a 1pccifjc inlerval of output llltc, stochastic models or models which have developed or by namtivc "slow".
IO II II 0DCeUary Chat tbi1 lllllbcmatical method . fuzzy COlllequen11 or fuzzy &Dlccedents is missing in
luch aystems. from mathematical logic. Tenn "slow" can h•vc diverse meaning if used ~y
llrictly ~ from the more familiar I · u
_
A general trouble in such constructed model is ho~ to different pcnons and must be understood with
~
loch 18 Boolean algd,ra. llglCI, -
1PPlications, -
moat of the an fuzzy logic solution
move from a given problem to a proper mathematical respect to the observed environment
- It i1 fiequcnuy lllggCaltd that the pow« of lhc h ~~ lranafonnation of a human solution. Also, fuzzy 'nl
modcl . CcrtaJ y,
today's advanced computer
. h Some values arc easy to categorize, while othen
brain iJ • flmction of ill ability to efficienu wnan ogic can IIIOdcl nonlinear fwictiona of uul'trary
A ...

technology makes it possible; however managmg sue

Y proceq COlllpl . can be difficult to determine bccauac of human
imprcclae information. cXJty to a dcairec! degree of accuracy.
'- systems is still too complex. undentanding ofdifferentsltuations.
- FL i1 a COOVcni t be simplified byt
- To introduce tbi1 'fuzzy logic" to a COl!lputationa1 en way to map an input space to an These complex systems can One can say "slow", while other can say ''not fast"
IClting, the critical clements are DOI DWDbcn but fuzzy • for a reasonable amoun
ou~ IJ)lee. FL is one of the toola uaed to IIIOdcl a employing a tolerance margm . during die when describing the same speed. These
letl. Dlul1i-1nput, DlUlti'-output 1ystem. of imprecision , vagueness and uncertainty dlfferencca can be distinguished with help of so-
modeling phase. called fuzz sets.
m Conatluctlon Management (SPPU) 9-9
lnlrO(lucUoll to Arttflclal lnteftlgence Technique r

Unlftdf of ~ : Range of all Po&siblc

zu....
1
,I: eonatn1ctlon Manapement (SPPU) . 0

9-10 , Introduction to Mi1k:1aJ lnteUlgence Tec:hnlq~ r

Uaually fuzzy logic control iy1tem i1 fonucd from
four major elementa repraented on Pig. 9.10.2
:fuzzification int.cdace, fuzzy infemicc engine.
value,, or concepta, applicable to a ay11e
variable.
111
rrlangll (x : a, b, c) •
{
(x-a)~(b-a)
{c-x)l(c-b) bsxsc
.:::b} Specified by three parametcn ( a,b,c) as follows:

bell (x: a. b,c) =

I
Iii
fuzzy rule mattix and defuzzificadon interfaoe.
_ When designing tho number of membership l'unctioru
for an , input variable. labela muat primarily be
0 X>C
I+!.=.£
1 I
Puz.ziticationi1 deaaibcd in dclail below.
detennioed for the membenhip functions. -t z. Trapezoidal MF Where the parameter b ia usually positive and we can
adjust c and a to vary the center and width of the
The number of labels relates to the number of rcgiona b
function and then use b to control the alopea.
iftplll 0~
that the wuvene should be divided, such that each label
definea I region of behavior. j o.~ .. ;... ,,..
A scope must be allotted to each membership function f o.e I \

that numerically recognizea the range of input values

that match up to a label.
j 0.4
0.2
0.8

0.8
,,
,,,
I
I \
\
\
I
\
I
·.
Fls,9.10.2 a
0.4
, I
I
The shape of the membership function should be 0 I \
0 20 40 eo 80 100 I
I \
\
9.11 Fuzzlftcatlon illustrllive of the variable. However this shape is also 0.2 __ , , ,I \

limiled by the computing iesources available. Fle, 9.11.2 ' ' , __

0
-t (SPPIJ-llay 14) Complicaled shapeS need more complex descriptive Specified by four parameters {a,b,c,d) as follows: -10 · -6 0 Ii 10
- Puzzification i1 the first llep in the fuzzy infemice equationa or luge lookup tables. Following figure Trapezoidal(x : a, b, c, d) _ (a) gian&1n& '•'
proceu. Thl1 contains a domain transformation where indicatea examples of possible shapes for membership
crisp inputa are transformed into fuzzy inputs. Crisp functiOD1. The most commonly ~ in practice are:
inputa are exact inputa IDCUIIMd by 9CIIIOl'S and puacd (x-a)~(b-a) aSxs'b
x<a }
into the control system for processing. IIICb 11 ..... al Ille llllmberllllp Function bSxSc
temperatme, prasure, etc. = { (d-x)j(d-c) cSxSd
• I ,
- Each crisp input that ii to be processed by the FlU bas 1. Trlanglar '-IF x:!:d - 0.4
ill own group of memberahip functiona CJC ICU to which
.. 3. GamsianMF
they are tranaformed. Thia group of membenhip 2. Trapezoidal MF .------- ,
functi0111 exiata within a univenc of dillCOWee that
3.BellahapedMF i 1
0
-10 . -6 0 Ii 10
hold& all 1elevant values that the crisp input can ! o.a 'c' (Contd-
po■aeal,
4.GBUl8ianMF f o.e
- The following &hows the llJUcture of membenhip t o.4 ,---;~:······ ..
functi0111 within I univenc of diJcoune fCJC a crisp Ii. Slgmoidal MF
! 0 .2 0.8
\
inpuL \
Fis, 0.3: Shapes olthe membenblp funcUom ,oo \
\
- Desree ol -'>enldp : Degree to which a crisp o.e I

I
I

~-,
.. 1. Trlanplar MF Fls, 9.11.3 \
value ia compatible to I memberahip function, 0.4
\
\
\
value from 0 to 1, also known II truth value CJC Specifie3 by two parameters {m,o) as follows: '
fuzzyinpuL
M-benblp llmctloa, MF: Definea a fuzzy 11C1
I 1
c,0.8 Gauasian (x : m, o)= ex"l -
J (x-m>2}
'i,2'· • .
0.2
0
-10 -6 0 5
,_

10
by mapping cri1p values from ita domain to the Q. (b) 'c'
fO,e Where m and o denote the center ~ ~idth of the
IICla IIIOCiated degree of membenhip.
jo.4 function, respectivelY, A small o will generate a "thin"
Crllp Input■ : Dillinct or exact inputa to a certain lo.2 MF, while a big o will lead to a ''flat" MF. o.a
1y11em variable, usually mcaa!Wd paramct.en
external from the control •)'Item. 00 .. 4. o.e
Label : Deacripave name uacd to identify a
membcnhip function.
.U·70JJMF 0.4

0.2

I
Scope : Or domain, the width of the membenhip Specified by three par&meten {a,b,c I as follows: O.I
0
o., -\0 -5 0 0 10
function, tho range of concepea, usually nwnben 0.2
over which a membership function i1 llllppCd. ' oo ao _, eo ., too (d) Cbangln1 '•' and 'b'

Fls,9.11,4 Fis, 9.11.5

m Coretiuction Manapement (SPPU)
~ S. Slpaoldal MF
9-11
lntroduCtlc!! to Ar11flcial lntell~ence Technique r
,W: Qon81ructlon Management (SPPU)
!!! ·Appllcatlona of Fuzzy Logic
9-12 Introduction to Altlllclal 1n11,:nce TIChn~ue r
- Data minina 1y11em1
, , ..,. + (SPPU- May 11) ~ " Chemical lnduatry
0.11 - Control of pH, drying,
0.11 Appllcatlona of Fimy Logic
- C!emical diatillllion (ll'O'CIICI,
0.4 - Polymer extnaion production, a cob oven gu cooling
0.5
1. Aeroepace plant
0.2
0 i..,.,i~------....;---1 2. Automotive
•s. Def- ·
-10 -5 0 5 10 - Underwater tarpt recognition,
(d)y,><y, 3. Bualness - Automatic target recognition of thermal inmml
Jl'li.9.11.6
F1g.9.11,7 imagea.
Specified by two paramelell {a, c J u follows: 4. Chelral lnduatry
- Naval deciaion support aids,
9.12 Characteristics-of Fuzzy Systems
Sipnoidal (x : a, c) =
l+e
la-cl 5. Defenae - Control of a hypervelocity inten:eptor,
Thele are two main characteristics of fuzzy systcDIS - Fuzzy set modelling of NATO decision making.
WIIIR c ia the center of the function and a control the 6.Elecll9nlc8
slope. Iha! give them beaer performance for specific applications: ~ '- Electroala
- Fuzzy syste111J are suitable for uncertain or 7. Financial - Control of automalic expoaun: in video cameru,
I
approximate reasoning, especially for the system with a - Humidity in a clean room,
I 8. lnclJlltrial
o.e I mathematical mo_del that is difficult to derive.
,
I Air conditioning lyale!IIJ,
0.8 Y,l, Fuzzy logic allows decision making with ~timated 9. Marine
-. Wuhing machine timing.
,,
I values UDder incomplete or uncertain informatj.on
Microwave ovcm,
,,
• 10. Medical
0.2
I 9.13 Advantages of Fuzzy Logic Systems - Vacmnncleanm.
0 ,, 11. Mining and Metal Proceeting
~ 7. FIDaDdal
- 10 ~ 0 5 10 - ~ c a l concepts within fuzzy reasoni~g are very
(a) Yi• 11&(:a:: 1, - 5), y2 • 11&(:a:: 2, 5) 12. Robotics - Banknote tranafer control,
11111ple.
_, Fund management,
- y 011 can modify 8 FLS by just adding or deleting rules 13. Securities
due to Oexibility of fuzzy logic. · - Stock mamt predicti0111.
Fig. C9.4: Applications of fuzzy lop: • 8. IDdultrial
0.8 - Fuzzy logic Systems can take imprecise, distorted,
noisy input information. - Cement kiln controll
~ 1, Aerospace
0.8
- FLSs are euy.to construct and understand. - Heat exchanaer control,
- Al~tude control of spacecraft,
0.4 - Activated sludge wutewater ttwment proc:eaa control,
- Fuzzy logic is 8 solution to complex problems in all - Satellite altitude control,
0.2 fields of life, including medicine, as it resembles Water purification plant control,
Flow and mixture regulation in almaft deciding
~1'::-0""'------~.___J human reasoning and decision making. Quantitative pearm analylia for lndumial quality
- vehicles. 1 •
~ 0 5 10 usurance,
9.14 Dlsa~antages of Fuzzy Logic ~ 2. Automotive
(b)IJ.-Jil Control of COllllrainl aatimclioo probleml In alnlCturll
Systams Trainable fuzzy systems for idle speed control,
deaign,
- , Automatic Gearboxes, Four-Wheel Steering
- Control of wala' purification plants.
- There is no systematic approach to fuzz
'I
\ designing. Y system Vehicle environment control ~ ,. MartDe
I - Shift scheduling method for automatic tnnamlss.ion,
~I Ys - They arc IIJldcrstandable only when simple.
- Autopilot for ahlps.
I - ' Intelligent highway _systems, - Optimal roulD aelectioo,
I
I - They arc suitable for the problems whi h d t eed
I high accuracy, C O DO n . - Traffic control, - Control of IUIOIIOIDOUI underwater vehiclea, ship
I
I
\ - Improving efficiency of automatic tranamluiOIII steering.
\
\
~ 3. ~ Bllllneal . . . . . Medical
0 5 10
- Decision-making support systems, - Medical diaplltic support ayatem,
(c) J1• 11&(1:: I, - 5), y, • 11&(:a:: -2, 5) - Personnel evaluation in a large company .
 --3i
lnl!Oduction to Artlflclal Intelligence Technlque r
~13 . , ¾ eoostnJCll<>n Management (SPPU) s-14 Introduction to Artiflclal Intelligence Technlque r
Control of arterial ;;..;;-during ~aeatlleaia,
• Reinforced conmte ~ construction and damage .
Genetic algorithms are a family of computational
Individual each solution vector which i1 each
Multivariable control of anacatheaia, • detection models in1piicd by evolution. These algorithms encode
_. Th■ f1■k1 of Hydrology & Water Reaource chromosome
a potential aolution to a specific problem on a simple
Modelling of DCUropalh logical findinga in Alzbeiiner'a
patients, ■ngln■■rlng chromosome-like data structure and apply Population total individuals
~ ,. . Radiology diagn01e1, _ For forecas~g rainfall, rainfall runoff, river stage, etc.
recombination operaton to these structures in rcapect
Population the number of cliroDlOSOme
of preserve critical inf0111l81ion.
Fuzzy infemice diagnosia of diabetea and prostale Hydrologic flow routing si7.c
_ Genetic algorithms are frequently viewed as function
cancer. Dam consllUCtion optimizer, even though the variety of problems to fimeas a function which evaluates how each
-t 11. Mlnlna 111d Metal Procealac - Surface W&Jr:r Hydrology which genetic algorithms have been applied are quite function solution suitable to the given task
- 1,,, Sinter plant control, - Url>an Water System wide. Phenotype expresaion type of' solution valuea in
- Decision making in metal fonning. - Filtration Optimization An implementation of genetic algorithm initiatea with a task world, for example, 'red,' "13
.. 11. Robotlai population of (typically random) chromosomes. cm", "4,.2 kg"
- Wllmhcd Management, Design of water distribution
One then evaluatca these structures and allocated
- Fuzzy control for flexible-link IDIDipulaton, system Genotype bit expression type of solution valuea
reproductive opportunities in such a way that these
- Robot ann control. - Hydrometcorology - uaed in OA search apace, for
chromosomes which represent a better solution to the
-t 13 SecurltJes example, "011," "01(01."
- Reaervoir Sedimentation Eatimatioo target problem are given more chances to 'reproduce'
than those chromosomes which are poorer solutiona. taking two fit individuala and then
- Dccisi9n systeDIJ for accuritiea trading. • Traffic ■nglnealng Breeding
- Fuzzy conccpta have its wi~ variety of applicalion for The 'goodness' of a solution is typically defmcd with intermingling there chromoaome to
For automatic control of traffic signals. regard to the current population. Basic Concepts of create new two individuals
the field of civil engineering. lntcgraling Fuzzy Logic
concepts wilil Genetic AlgorilhDIJ (OA) ~ Neural - traffic IICcident .estimation and traffic control system Genetic Algorithms are
Mutation changing a random gene in an
Networks (NN), the complex problems can be - road Construction Genetic algorithms are the main paradigm of individual
efficiently and effectively solved in order to arrive at - Geo-mechanical and PavC111C11t Systems evolutionary computing. Genetic Algorithms are
?Ptimal solution. Thcae conccpta are effectively Ulcd . _. inspired by Darwin's theory about evolution - the
R■ll■blllty of .-truclurM "survival of the fittest". In nature, competition among
9.17 A Simple Genetic Algorithm
m:
For damage aaeaament in structures individuals for scanty resources results in the fittest ♦ (SPPU • May 12)
er Structural ■naly■I■ ■nc1 o.■lgn individuals dominating over the weaker ones.
- Metal atructurN The mechanics of a simple genetic algorithm are
- For 1tructura1 optimization 111d optimum De&i of - Genetic Algorithms are the methods of solving ·
structure,, go For predicting fatigue 111d Creep characteristics surprisingly simple, involving nothing mare complex
problems by mimicking proceases nature uses; ie.
Ar Otlwtl■lda Selection, Crosses over, Mutation and Accepting, to than _copying strings and exchanging partial strings.
- Computation Motphogeneaia of Diacretc llnx:turea.
Material behaviour modelling. · develop a solution to a problem. Simplicity of opezation and power of effect are two of
Oeotcchnica! engineering
- Genetic Algorithms are adaptive heuristic search based the main atttactions of the genetic algorithm approach.
- Structural Damage UICUmcnL F.nvironmenta1 F.ngineering and Oeosciencca on the evolutionary ideas of nalUral selection and
- FJnitc Element Analyaia. Fluid dynamic,, fluid mechanics and heat transfer A simple genetic algorithm that yields good results in
genetics.
.., Con■tructlon fleld - Genetic Algorithms are intelligent exploitation of many practical problems is composed of three
.9.18 Genetic Algorlthma operators.
For management problema like CODllruclion l!Chcdu1in random search used in optimization problems.
of the project · g -t (SPPU. May 12. Oct. 12, May 13, May 14, May 18) - Genetic Algorithms, although randomized. exploit
OperalOrll In GerMtlc Algortlhm
• Introduction historical information to direct the search into the
- For planning of life cycle of project ptobloma like
region of better performance within the search space.
selection of beat construction equipment - Genetic algorilhDIJ initiated and developed in early 1.Repf0CIUctlon
- Project management CODllluction technology.
Material management
1
~°'. ~ John Holland are unorthodox eearch and 9.16.1 Technlcal Term■ uted In Genetic
Algorithm 2. er-er
optimizatjon algorilhDIJ, which mimic IOIDC of the
Predicting Construction Material Pricea. PIOOealea of nat11n11 evolution.
vector which represents soludona of 3. Mutation
project evaluation, diagnoaia, decilioD-makin and - Genetic algorilhDIJ carry out ~ Chromosome
through . random eearchea application task
prediction, g . a given set of allernalivea with the . f F1a, C9.5 : Operator of In poetic alaortthm
finding the beat al · · 8IJD O
Gene each solution which consists of a
Building design and optimization. of goodneaa, • tcrnative wtth respect to given criteria
chromosome
bridge oonstrucdon - Theae criteria are
objecti ~ulrcd to be expreucd in of Selection choosing parents' or offspring's
Con=tc mix design. 00IIClele maiglil.eatimaaion
ID
fitneaa run: function
OIIJ.
which i1 usually refemd to u
terms
chromosomes for the next generation
 Introduction 10 ArtlflClal Intelligence Technl~ ~ c;onitruotion Management (SPPU) ~18
~~Manageme nt (SPPU) 9-15 lnlroduollon lo At1111cia1 lntelllotnct Teohnlque r
-+ 9,17.1 Reproduction -+ 9.17.2 Cr011over ;;,;.,a radons. A one aite cl'Ol80ver operator i1 executed by llrina, there lJ problbiUty that each bit 1111)' become
opedoffllY aelecting a crouing aite along the 1tring and IIIUtlled.
Reproduction is a i-oceaa in which individual strings _ To recomblne two string• to acquire a better string 1
crosaover operator is used.
ran
by CllChanging all bits on the right aide of the
croulng _
Thia probability ia UIUallylOllleWhlta email value,
are copied according to their objective function values, . ahown in Fig. 9.17, l.
In Cl'OIIOVer opcratlon, recombinatlon Pl'OCeaa 11te81 called u llllllation probability p..
f biologill call thia flmction the fitncsa function. _
nerates dlssimllar indlviduals in the aucce51ivc In one aite crooover, a crouover alte la chosen A coin toll mechani1m l1 employed; If random number
Intuitively, it can be consider function f u some :neratlons by uniting material from two individuals or - domlY which l1 1hown u vertical Unea. between uro and one lJ laa than Ille lldllion
measure of profit, utility or goodneaa that need to the previous generation. ran portion right of the selected lite of these two probability, then the bit ia lnvated, 10 dllt ~
muimize. of ltrinp.. The
_ In reproduction, good strings in a population are
- n,e "'w exchanged to form a new pair
string• ·- becomea one and one becomea um.
- Copying strings according to their fitness values means probabilistically allocated a larger number of copies
· a• are thua an amalgamation of the old airings.
new 1tr1n.,.. _ Thia hell" in inlrodacins a bit al divcnity lo lbe
that strings by means of higher values have higher and I mating pool is formed while no new strings are Two llw.•• crosaover ii a variation of. the one aite populalion by ICllfering lbe oa:uional poillll.
probability of conlributing one or more offspring in formed in the reproduction phue. - croaaover , except that two cro11over 11te1 are chosen - Thia random atlaing would iault la bea.er optima. OI
next generation.
_ In the crossover operator, new strings are formed by and the bits between the aitea are exchanged. even modify a part of lffldic code 1h11 will be
- This opentor, certainly, is an artificial version of switching information among strings of the mating 011111001 bend"ICIII la Iller opmtiona.
011101 String l
natural aelcction, a Darwinian swvival of fitness among String 1
pool. 100 - On the other band, it might prnducc a weak mmal
string cre&IUJu.
that will never be 11elec:led forfardlcropenlicm.
- In the crossover operation, the two strings which are String 2 0lljOllOO
- In naturl) populations fitness is determined by a String 2 110111 - The need for IJlulllion is to ~ I poml ill lbe
parlicipating are known as parent strings and the
=atun:'s ability to survive im,dators, pestilence, and 001 neighborhood al die CUl'fflll point. tbmby aaaiaillc •
resulting slringa are known as children strings.
the other obstacles to adult hood and subsequent _ After crouover
Before crooover local 1Qn:b llllUlld 1be cmmit IOlmion.
reproduction. In our unabubcdJy artificial setling. the It is intuitive from this construction that good sub-
strings from parent strings can be combined to produce (a) One lllte ~ver operation - The mlllllioa is aJ,o uaed to maill&aiD civtnily iD ~
objective function is final arbiter of siring creature's
life or death. a better child siring, if an appropriate site is selected. populalioa. For example. lbe f ~ popalalioe
String 1 011111"° having fOID' eight bit lllrinp may be aJlllidcnd:;
- Reproduction or selection is usually the lint operator - Witli a !IDdom site, the children airings formed may or String 011!011!00
may not have a combination of good sub-strings from 01101011
applied to a population. Reproduction picks out good 1
airings in a population and forms a mating pool. parent airings, depending on whether or not the String 2 0111011101 00111101
String 11011101<>1
- The crucial idea is that above-average strings are crossing site falls in the appropriate place. (0)10110
2
picked from the cwrent population and duplicates of - But this is not a matter of serious concern, because if After crossover 01111100.
them are inserted in the mating pool. .Before croaaover
good strings are produced by crossover, there will be IIIJlicaf that all fa. 11rinp baft IO ia die left
- It an be . ....,. mqam:a
- The commonly used reproduction opentor is the more copies of them in the next mating pool generated (b)Two lllte croaover opendloo
IIIIJll bit p0liliml. If die true opllllllllD . ..
by crossover. Fig. 9.17.1
proportionate selection operator, where a lllring in lhe
1 iD lhal poailioe. lbell - : :~ 10
current population is =fully chosen with probability - Hence, to picserve some of the good strings that cnmover opmror cb:libcd
~ 9.17.3 Mutation
proportional to the siring' I fitnea. already exist in the mating pool, all slrings in the aare 1 iD 1h11 p0liliml.
_ Mutalion adds new informatlOII · a random way to the
. . ID .
- Thua, the i~g in the population is selected with mating pool are not used in crossover. ........ . ..__.,_ oC IIIIIIIDml iDllodaca pnlbmililJ pa o{
. d eventuallyaids to av01d _ .... me...... aimplead
probability proportional to fi.Since the population siu - When a crossover probability pc is used, only lOOpc genetic search process 811
tuming O iD10 l.Tlae dne openran an:
is usually kept fixed in a simple GA, the cumulative per cent strings in the population are used in the getting trapped at local optima. . . ......,orwanl
. that introduces divenity ID the ltnJ_, . and die
probability for all siring in the population must be croasover operation and 100(1-pc) per cent of the - It II an opemlOr . to bcaJme
The , - . I -
" _ opml« choolel good llllllp
IUb-tlmP faa
one.Therefore, the probability for selectingi111string is, population remaina as they are in the cunent population whenever the population tends . and - na>mDlla _ .
population, aoucmr opemar ,ccoo,binea _.- .
N homogeneous due to repeated use of reproductiOII
. •-to-•bellr:r ........
F, = Lr. - A crossover operator is mainly responsible for the crosaover.operaton.
-
goodatrmp___.,
•
--
opml« alias a arriag locally ap,caDI 8
.
i =I search of new strings. There are various crossover - Mutation is used to mamtam .,~u- .
. . --,;c diversityalfrom·tbm
f genelJC gon
one
The mulalJOil o{ dl£,ae claiml -
Where, N is the population sil.C. One way to attain this OJ)Clllora exist in the genetic algorithm literature.
.
generation of a population ° the beUer alriag. E.ai lboagll - w!Ji)eamiag&llmg.
. ·• . .
proportionate selection is to use a roulette-wheel with hromosomea to the next and it may cauae ~ ad/or - i will be
- Moat commonly adopted are one site crossover and two C dissimilar from thole if bad aliiDp an: a - ' ~
the circumference marted for each siring proportionate chromosomes of individuals to be expected 1118' odaclian opml« ill die ac:n
site crossover. In moat crossover operaton, two airings elimi,.-1 by the iep
to the airing' I fitneas. of their parent individuals. .
are chosen from the mating pool at random and some . 111d if good aliiDp - aeatal. Ibey will be
diJIUrt,ing genellC
portion of the slrings areawitched between the airings. - Mutation is the process of ~ y . .. ~ the bill g,ent:rlllllil .
. . Th ,,,_.,.,,, 11 the bit )eve1• w,_, mcraainglyeq,Pmittd
- At siring level crossover Operation is done using mformabon. ey -.-- . to the new
rando are being copied from the cum:nt IIIMg
m1y selecting two strings for croasover
M Conatructlon Manie::= (SPPU) lnt,oduetlon 10 Artlflcial lntelllgence Ttchnlqu11
w: eontWO!lon Manae=t (SFIPU)
Application of the,e open!On on the current
e,17
-- .. 6. Ecoloalcal model■
OA• hu been uaed to model ecological phenomena
9-18

➔ L
lntl0ductlon to Al1lllclal lm:'!P:':: Ttchnlque r
Ac:ce.lblllt7
populllion CTellel a new population. Thia new· Exert ayatema 1R1 alway1 available for 111C when human
population la llled to generate aublequent populallona ,uch 81 biological arma l'IICCI, hoat-paruite co- expert■ are not readily available.
and 10 on, yielding 10luti0111 that arc cloaer to the 1. Optimization evolutlona, symbloel1 and reaource owe in ecologlea. .. IL Comllteacy
• optimum aolutlon. -t 7. population pnetia modell
2. Automatic programming Expert 1yatems are lea■ likely to contain inacancica
_. 11 'The valuea of the objective lunction of lhe lndividual1. OAJ h81 been uaed to atudy queation1 In population provided lhe expert 1yatem bu good knowJedp
< of the new populllion an, apln detennlned by 3. Machine and robot IMmlng genetlc1, auch 81 "under what condition■ will a gene for ~talion In IICCUl'lciea or erron can be eully
recombination be evolutionarily viable?' Interactions prevented
decodlna the atrinp. The8e valuea expreaa the fitneu
4. eoonomlc ~ between evolution and learning: OAI bu been Uled to
of the aolutlona of the new gencrali0111. ➔ UL 11me coaatnlnta
atudY how individual learning and 1pecica evolution
- Thia complete, one cycle of genetic algorithm called • 5. Immune ayattm modlll The numben of copica or an expert 1yatem can be
affect one another. . made wbereu, the training prnceu of human expert la
generation. In each generation if the aolution i1
8, EcologJcalmodtl• ,. 8. Model■ or IOdal ayateim time COlllllming and expe111lve.
Improved, It la atored u the beat aolutlon. Thia ia
OAI baa been uaed to study evolutionary upecta of ➔ Iv. Stability
repeated till convet'lence, 7, P0jdatlon genetlce moclefa
aoclal 1y1tems, such u the evolution of cooperation,
It can aui11 1 human expert in problem aolvlna and i1
8.18 Appllcatlon1 of Genetic Algorithm e. Modell of aoclal ayatema the evolution of communication, and trail-following
more likely to conaider all p0111lrilitica.
behaviour in anta.
- An effective Genetic algorithm ~preaentation and ➔ '. Effldenc7
Fl&, CU: Appllcatlona or GA Expert ayatem i1 capable of ~viewing all the
murungful fitneaa evaluation are the keys of the 9.19 Expert 1y1tem
.. 1, OpUmbatlon tranaacti0111 u compmed to human expert■•
auoceu In Genetic algorithm applicllions.
In an artificial intelligence, an expert 1y1tem i1 a
- The enlrelly of Genetic algorithms come& from their OAI have been uJed in a wide variety of optimization 9.19.3 Umltatlona of Expert System
computer 1yatem that emulatea the decilion malcing ability
limpllcity and elegance u robuat search algorithms u wb, including numerical optimization, and
of human expert. An expert system is uaed to extnct the (i) Limitatiooa of the ICIChnology
well u from their power to find out good aolutions combinatorial optimization problems such u ·traveling
information of human expert within specific domain and
quickly for difficult hipHtimeualonal problema. lalcaman problem (TSP), circuit design, job shop (ii) Difficult knowledge acquisition
makcl this knowledge available to leas experienced uses
• Oenetlo algorfthme are UNful and acbeduling and video & sound quality optimization. (ill) Expert 1y1tems me difficult to maintain
efficient through computers codes program.
whan .. 2. Automatic: Pro&rammlnl (iv) Align development coata.
OAI hu been uJed to evolve computer programs for 9.19.1 Characteristics of Expert System
l. The 1C11Cb space is large, complex or poorly 9.19.4
specific tasb, and to design other computational Appllcatlons of Expert System In Civil
undentood (i) High performance Engineering
structure&, for example, cellular automata and sorting
2. Domain knowledge i1 acarce or expert knowledge is networb.
(ii) Undentandable
difficult to encode to narrow the aearch apace (iii) Reliable
i) In environmental civil Engineaiq. one of the 11101t
. . 3. Machlae and robot learnlq important clwactfflllics ii the involvement of
3. No lllllhcmalical analysi1 ii acceuible (iv) Highly reaponsive
OAs has been U8ed for many machine- learning multidiaciplinary areu (chemiatry, biology, Ouid
4. Traditional aean:b methoda be unaucceaaful appliclliona, including cluaification and prediction, 9.19.2 Advantagn of expert system mechanics, l!Lllhemllica, atatiatics, economics. IIIIC
and protein &true~ prediction. GAs have also been
- The advantage of the OA approach ii the eaae with Aclvantagn ol expelt ayatem etc). Expert system in tbla na 111ually deal with
uJed to design neural networks, to evolve rules for
which it can handle arbitrary types of comtrainta and hazardous waste management, waler quality procection
learning classifier systems or symbolic production
objectivea; all 1uch things can be handled u weighted systems, and to design and control robots. I. Acceaslblllly and water resources,
conatituenta of the fitDcU function, making it euy to ➔ 4. Economic models ii) In stNCtWal engineering, expert 1yatema have aided
adapt the OA acheduler to the particular necessities of a H. Cona181811CY design, analysis, inapcction and maintenanee and code
OAI bu been U8ed to model processes of iMOVation,
very wide range of pouible overall objcctivca. checking. There is also • great poeential applicllion for
the development of bidding strategies, and the UI. Time OOMtralnta
emergence of economic markets. diagnoail and interpretali0111 problema. Molll of the
- OAa bu .been used for problem-aolving and for
Iv. StablMty expert 1y1tema developed in thla na involve
modeling. Genetic algorithms arc applied to many ➔ s. Immune ayatan modela
interaction wilh conventional calculua and analysi1
acientific, engineering problema, in buaineu and OAI has been U8ed to model virioua upecta of the v. Elllc:lenCY · programl, and the expert ayatem ii an aid to conceive
entcrtallllDCllt, comprising : natural immune system, including somatic mutation models and to undentand atructunl behavior. Expert
F11, C9.7 : Advmtaaes or espert llydem
during an individual's lifetime and the discovery of iyatem 11 uJed for diapoai1 of safety and Integrity of
multi-gene families during evolutionary time.
atructunl aystema, diagnOll1 of buildin~ defccta.
 ~
~ Construction Manapen:nt (SPPU) 9-19
1
n
~ to Artificial Intelligence Technl
~
M construction Manaaement ISPPUI 9-20 lntroductlon to Allfflc!al lnteHIQence Technlaue r

Q.8 How genetlc algorithm can be appl~ · to

iii) In geotechnical engineering there are also many _. 2. J)lapoos
applica!iooa involving dugno,li• and t,eatmCDt of dam construction scheduling wtth or without reeourcee
_. (a) Failure dlagDOSls
seq,age problema, analysis of geotcchnical Wdslidcs, rockslidcs, building failure, water lnd a.,1 Define Artlflcfal Intelligence. constraints. (Sectfona 9.111 and 9.18) (8 Marb)
investigation problems, soil classification and inference sewage ttea1D1ent p Iant f81·1we, - failure of a.2 Write a short note on : Q. 7 Explain with example fuzz1 logic appllcetlon In cMI
of shear lllreaaca 111ing cone penetration teSt data. englneerfng.(Sectlon 9.18) (8 11■111■)
(I) Artlflclal Neural Netwolk
choice of alternative designs for retaining wallJ etc. consuuction schedules etc all require collecting
(11) Neural Netwolk Architectures
a. a How Artlflcfal Neural NetwOlk helpe 1n m■ter1a1
iv) •Tranaportation engineering expert systems have been a and evaluating data of various degrees of CCrtain1y management In clvll engineering .(8 ~.
'
1
' very useful alternative in planning, designt contt0l and to identify the most likely cause of failure. . (Ill) Blologlcal Neural Netwolk
operation, management and maintenance activities. Remedial dlap!Olil (Iv) Fuzzy Logic System
May2013 :;1> o
Expert ayatem helps in design of highway noise .. (b)
,... ,barrirn. generate 11'11ffic control plans to avoid The expert system helps in diagnosis of eltisting a.3 'Explain wolklng prtnclple of artlflclal neuron. Q. 9 What do you understand by Artlflclal Intelligence?
And. write ltB applk:atlona In Conatructlon
congealing Bitualion, evallllling concrete pavement civil engineering systems to determine potential .a. 4 Explain briefly the terms cell body, axon, synapse,
management.(Sectlond.1 ■ncl9.18) (12 IQib))
performance e.g. topping and management of low dendrite and neuron with reference to a biological
failure and dysfunctions.
volume flexible pavement that recommends pavement neural netwolk. 0. 10 Explain the analogy between· Bk>loglcal neuron and
n"1abilltation and maintenance strategies. Monitoring Artlflclal neuron. (Seetlona 9.2 and 11.8) (4 11■111■)
_. (a) Performance Monitoring
a. 5 What are the features of Artlflcfal Neural Netwolk
Q. 11 What are the appllcetlona of ArtlflcfaJ Neural
9.19.6 Appllcatlon1 of Expert 1y1tem In and Biological Neural Netwolk 7
networks In construction management 7
construction With the increasingly available m i ~ a.6 Ust and briefly describe the tennlnology of ArtJllclal
and sensors providing input to expert systems, ~ Neural Netwolk.
(Section 9.18) (8 Marb)"
Applloatlona ol Expert apwn In
canalnlcllon time monitoring and time dependent diagnosis a. 7 Explain fuzzy logic and Its application In clvll Q. 12 What la an Expert system 7 How ArtlflcfaJ
lntefllgence Is useful In the development of Expert
may be performed on structures, foundations, engineering.
system. Explain wtth suitable e ~le.
1. l;xlatlng condlllons
ieservoir systems etc. as well as construction Q. a Discuss different membership function In Fuzzy
(Section 9.1) (1011■111■)
~uipmenl Logic.
2. Dlagnou
~ (b) Procas monitoring a.9 What are various appllcetlons of Fuzzl Logic? May2014
(a) Falll.n diagnosis
Monitoring of design and CODSl:n\Ction processes
a. 1o Describe the technlcel terms used In Genetic a. 13 Discuss appllcetions of ANN In CMI Englneenng.
Algor1thm (Section 9.18) (8 Marica)
(b) Remedial dlagnoela to control costs and duration could be provided by
a. 11 Explain Genetic Algorithm and its appllcetlon. Q. 14 What arc the different appllcetlon of Arttflclal
3.Monllodng expert systems inco!porating knowledge about
Intelligence In Construction Management.
attentive schemes, reaction to contingencies etc•. 9.20 University Questions and Answers (8 Marie■)
Planning Q. 15 What Is forward and backward propagation In ANN.
M11y2012
(b) Proceaa monitoring ~ (a) Project planning (Section 9.3) (8_Marb)
Q. 1 Define Artlflclal Intelligence and write down Q. 18 What Is fuzzy Logic 7 Olscusa dlffetenl membership
Bxpert system helps in planning of design and
appllcetions of Artificial Intelligence In CMI function In Fuzzy Logic.
construction projccta where there is a large range
9!1glneerfng. (Sectfona 9.1 and 9.18) (12 Marb)
(a) Projact p1amng of potential actions, many conflicting (Sectfona9.9 ancl9.11) (8 Marb)
0.2 Differentiate between Blologlcel neuron and artlflcfal
miuirements and possible contingencies to
(b) Maao-plaMing neuron. May201s
consider.
(Sectfona9.2 and 9.8) (4M■rb)
5.DNlgn ~ (b) Macro-planning Q . 9(a) Explain In detail Importance of artJflclal Intelligence
0. 3 · Write short note on the following : (12 M■rb) In Infrastructure projects.
Fla- C9.8 : ApplJcatlom of ezpen ryllem In COllltrucdoo '.IBMing of large capital projects such sa
(I) Artlflcfal Neural netwolks. (Section 9.2) (Ans. : Refer ssctJon 9.1) (8 Marb)
mfraatructure rehabilitation plans, could similarly
~ 1. EDICiaa cundldom (ii) Fuzzy Logic. (Section 9.8) Q. 9(b) Define neural network. What Is forward and
be assisted by eitpert systems, especially where
:111° gathering and inlelprelalion of exillling conditions public reaction and preferences strongly influence backwanl propagation In ANN ?
(Ill) Genetic Algorithm. (Section 9.17)
11 the lint atep of any civil engineering design. Expert theplana. (Ans. : Relrlr s«tiona 9.2 lltld 9.3) (8 Marb)
0. 4 What la training of neurons ? Explain.
ayatem used f°'." inte,pretation of atructural condition ~ 5. Design Q. 9(c) What are the appllcetlona of fuzzy logic.
and load capacity of structure baaed on observati · (Section■ 9.2 and 9.3) (4 Marke) (Ans.: RBfersBCtlon9.15) (4 Marb)
and non-destructive teata in the abaence of adeq ona Bxpert syatem helps for selection of initial design Oct2012 Q. 10(a) Write short note on : Genetlc Algorithm
J)arlmeten, for the
dulgn documenta interpretation of tnffic conditi: . . evaluauon
.
and solutions, ,or , (Ans. : Refer HClion 9.16) (8 Marb)
and demand& for lranfportation improvemepta etc. modificstion and Rdca. 0. 5 Explain application of artlflcfal neural network In clvll
•gn of unsatisfactory components elC- englneer1ng.(Sectlon 11.18) (8 Marb)