MB0048 Operation Research Set-1

Q1. a. Explain how and why Operation Research exec#ti!e decisions. ethods ha!e "een !al#a"le in aidin$ a'in$ process and the role o&

". %isc#ss the #se&#lness o& Operation Research in decision co p#ters in this &ield. (ns.

Churchman, Aackoff and Aruoff defined Operations Research as: the application of scientific methods, techniques and tools to operation of a system with optimum solutions to the problems, where optimum refers to the best possible alternative. The objective of Operations Research is to provide a scientific basis to the decision-makers for solving problems involving interaction of various components of the organisation. You can achieve this by employing a team of scientists from different disciplines, to work together for finding the best possible solution in the interest of the organisation as a whole. The solution thus obtained is known as an optimal decision. You can also define Operations Research as The use of scientific methods to provide criteria for decisions regarding man, machine, and systems involving repetitive operations.OR Operation Techni!ues is a bunch of mathematical techni!ues." ". )Operation Research is an aid &or the exec#ti!e in a'in$ his decisions "ased on scienti&ic ethods analysis*. %isc#ss the a"o!e state ent in "rie&. (ns. Operation Research is an aid for the e#ecutive in makin$ his decisions based on scientific methods anal%sis". %isc#ssion+An% problem, simple or complicated, can use OR techni!ues to find the best possible solution. This section will e#plain the scope of OR b% seein$ its application in various fields of ever%da% life. i& ,n %e&ense Operations+ 'n modern warfare, the defense operations are carried out b% three major independent components namel% Air (orce, Arm% and )av%. The activities in each of these components can be further divided in four sub*components namel%: administration, intelli$ence, operations and trainin$ and suppl%. The applications of modern warfare techni!ues in each of the components of militar% or$anisations re!uire e#pertise knowled$e in respective

fields. (urthermore, each component works to drive ma#imum $ains from its operations and there is alwa%s a possibilit% that the strate$% beneficial to one component ma% be unfeasible for another component. Thus in defense operations, there is a re!uirement to co*ordinate the activities of various components, which $ives ma#imum benefit to the or$anisation as a whole, havin$ ma#imum use of the individual components. A team of scientists from various disciplines come to$ether to stud% the strate$ies of different components. After appropriate anal%sis of the various courses of actions, the team selects the best course of action, known as the +optimum strate$%,. ii& ,n ,nd#stry+ The s%stem of modern industries is so comple# that the optimum point of operation in its various components cannot be intuitivel% jud$ed b% an individual. The business environment is alwa%s chan$in$ and an% decision useful at one time ma% not be so $ood some time later. There is alwa%s a need to check the validit% of decisions continuousl% a$ainst the situations. The industrial revolution with increased division of labour and introduction of mana$ement responsibilities has made each component an independent unit havin$ their own $oals. (or e#ample: production department minimises the cost of production but ma#imise output. -arketin$ department ma#imises the output, but minimises cost of unit sales. (inance department tries to optimise the capital investment and personnel department appoints $ood people at minimum cost. Thus each department plans its own objectives and all these objectives of various department or components come to conflict with one another and ma% not a$ree to the overall objectives of the or$anisation. The application of OR techni!ues helps in overcomin$ this difficult% b% inte$ratin$ the diversified activities of various components to serve the interest of the or$anisation as a whole efficientl%. OR methods in industr% can be applied in the fields of production, inventor% controls and marketin$, purchasin$, transportation and competitive strate$ies. iii& -lannin$+ 'n modern times, it has become necessar% for ever% $overnment to have careful plannin$, for economic development of the countr%. OR techni!ues can be fruitfull% applied to ma#imise the per capita income, with minimum sacrifice and time. A $overnment can thus use OR for framin$ future economic and social policies. iv& ($ric#lt#re+ .ith increase in population, there is a need to increase a$riculture output. /ut this cannot be done arbitraril%. There are several restrictions. 0ence the need to determine a course of action servin$ the best under the $iven restrictions. You can solve this problem b% appl%in$ OR techni!ues. v& ,n .ospitals+ OR methods can solve waitin$ problems in out*patient department of bi$ hospitals and administrative problems of the hospital or$anisations. vi& ,n /ransport+ You can appl% different OR methods to re$ulate the arrival of trains and processin$ times minimise the passen$ers waitin$ time and reduce con$estion, formulate suitable transportation polic%, thereb% reducin$ the costs and time of trans*shipment. vii& Research and %e!elop ent+ You can appl% OR methodolo$ies in the field of R12 for several purposes, such as to control and plan product introductions.

Q0. Explain how the linear pro$ra in$ techni1#e can "e help&#l in decision- a'in$ in the areas o& Mar'etin$ and 2inance. (ns. 3inear pro$rammin$ problems are a special class of mathematical pro$rammin$ problems for which the objective function and all constraints are linear. A classic e#ample of the application of linear pro$rammin$ is the ma#imi4ation of profits $iven various production or cost constraints. 3inear pro$rammin$ can be applied to a variet% of business problems, such as marketin$ mi# determination, financial decision makin$, production schedulin$, workforce assi$nment, and resource blendin$. 5uch problems are $enerall% solved usin$ the simple# method." ME%,( SE3E4/,O5 -ROB3EM. The local Chamber of Commerce periodicall% sponsors public service seminars and pro$rams. 6romotional plans are under wa% for this %ear,s pro$ram. Advertisin$ alternatives include television, radio, and newspaper. Audience estimates, costs, and ma#imum media usa$e limitations are shown in 7#hibit 8. 'f the promotional bud$et is limited to 98:,;<<, how man% commercial messa$es should be run on each medium to ma#imi4e total audience contact= 3inear pro$rammin$ can find the answer. Q6. a. .ow do yo# reco$nise opti ality in the si plex ". 8rite the role o& pi!ot ele ent in si plex ta"le7 (ns. 5imple# method is used for solvin$ 3inear pro$rammin$ problem especiall% when more than two variables are involved S,M-3E9 ME/.O% 1. Set #p the pro"le . ethod7

That is, write the objective function and the constraints. 0. 4on!ert the ine1#alities into e1#ations.

This is done b% addin$ one slack variable for each ine!ualit%. 6. 4onstr#ct the initial si plex ta"lea#. .rite the objective function as the bottom row. 4. /he ost ne$ati!e entry in the "otto row identi&ies a col# n.

:. 4alc#late the 1#otients. /he s allest 1#otient identi&ies a row. /he ele ent in the intersection o& the col# n identi&ied in step 4 and the row identi&ied in this step is identi&ied as the pi!ot ele ent. The !uotients are computed b% dividin$ the far ri$ht column b% the identified column in step >. A !uotient that is a 4ero, or a ne$ative number, or that has a 4ero in the denominator, is i$nored. ;. -er&or pi!otin$ to a'e all other entries in this col# n <ero.

This is done the same wa% as we did with the ?auss*@ordan method. =. 8hen there are no ore ne$ati!e entries in the "otto otherwise> we start a$ain &ro step 4. 8. Read o&& yo#r answers. row> we are &inished?

?et the variables usin$ the columns with 8 and <s. All other variables are 4ero. The ma#imum value %ou are lookin$ for appears in the bottom ri$ht hand corner. Example )iki holds two part*time jobs, @ob ' and @ob ''. 5he never wants to work more than a total of 8; hours a week. 5he has determined that for ever% hour she works at @ob ', she needs ; hours of preparation time, and for ever% hour she works at @ob '', she needs one hour of preparation time, and she cannot spend more than 8A hours for preparation. 'f she makes 9>< an hour at @ob ', and 9B< an hour at @ob '', how man% hours should she work per week at each job to ma#imi4e her income= Solution: 'n solvin$ this problem, we will follow the al$orithm listed above. 1.Set #p the pro"le . That is, write the objective function and the constraints. 5ince the simple# method is used for problems that consist of man% variables, it is not practical to use the variables #, %, 4 etc. .e use the s%mbols #8, #;, #B, and so on. 3et #8 C The number of hours per week )iki will work at @ob '. and #; C The number of hours per week )iki will work at @ob ''.

't is customar% to choose the variable that is to be ma#imi4ed as D. The problem is formulated the same wa% as we did in the last chapter. Maxi i<e S#"@ect to+ ;E8 F #; G 8A #8 H <I #; H < 0. 4on!ert the ine1#alities into e1#ations. This is done b% addin$ one slack variable for each ine!ualit%. (or e#ample to convert the ine!ualit% #8 F #; G 8; into an e!uation, we add a non*ne$ative variable %8, and we $et #8 F #; F %8 C 8; 0ere the variable %8 picks up the slack, and it represents the amount b% which #8 F #; falls short of 8;. 'n this problem, if )iki works fewer that 8; hours, sa% 8<, then %8 is ;. 3ater when we read off the final solution from the simple# table, the values of the slack variables will identif% the unused amounts. .e can even rewrite the objective function D C ><E8 F B<E; as J ><E8 J B<E; F D C <. After addin$ the slack variables, our problem reads Objective function: J ><E8 J B<E; F D C < D C ><E8 F B<E; #8 F #; G 8;

5ubject to constraints: #8 F #; F %8 C 8; ;E8 F #; F %; C 8A #8 H <I #; H < 6. 4onstr#ct the initial si plex ta"lea#. .rite the objective function as the bottom row. )ow that the ine!ualities are converted into e!uations, we can represent the problem into an au$mented matri# called the initial simple# tableau as follows. #8 #; 8 8 ; 8 J>< JB< %8 8 < < %; < 8 < D < < 8 C 8; 8A <

0ere the vertical line separates the left hand side of the e!uations from the ri$ht side. The hori4ontal line separates the constraints from the objective function. The ri$ht side of the e!uation is represented b% the column C. The reader needs to observe that the last four columns of this matri# look like the final matri# for the solution of a s%stem of e!uations. 'f we arbitraril% choose #8 C < and #; C <, we $et .hich reads %8 C 8; %; C 8A D C< The solution obtained b% arbitraril% assi$nin$ values to some variables and then solvin$ for the remainin$ variables is called the "asic sol#tion associated with the tableau. 5o the above solution is the basic solution associated with the initial simple# tableau. .e can label the basic solution variable in the ri$ht of the last column as shown in the table below. #8 #; 8 8 ; 8 J>< JB< 4. /he %8 8 < < %; < 8 < D < < 8 8; 8A < %8 %; D

ost ne$ati!e entry in the "otto

row identi&ies a col# n.

The most ne$ative entr% in the bottom row is J><, therefore the column 8 is identified. #8 8 ; J>< #; 8 8 JB< %8 %; 8 < < 8 < < D < < 8 8; %8 8A %; < D

Q4. 8hat is the si$ni&icance o& d#ality theory o& linear pro$ra r#les &or writin$ the d#al o& a linear pro$ra in$ pro"le .

in$7 %escri"e the $eneral

(ns.3inear pro$ra in$ K3-& is a mathematical method for determinin$ a wa% to achieve the best outcome Ksuch as ma#imum profit or lowest cost& in a $iven mathematical model for some

list of re!uirements represented as linear relationships. 3inear pro$rammin$ is a specific case of mathematical pro$rammin$. -ore formall%, linear pro$rammin$ is a techni!ue for the optimi4ation of a linear objective function, subject to linear e!ualit% and linear ine!ualit% constraints. ?iven a pol%tope and a real* valued affine function defined on this pol%tope, a linear pro$rammin$ method will find a point on the pol%tope where this function has the smallest Kor lar$est& value if such point e#ists, b% searchin$ throu$h the pol%tope vertices. 3inear pro$rams are problems that can be e#pressed in canonical form: where x represents the vector of variables Kto be determined&, c and " are vectors of Kknown& coefficients and is a Kknown& matri# of coefficients. The e#pression to be ma#imi4ed or minimi4ed is called the ob!ective function KcTx in this case&. The e!uations x G " are the constraints which specif% a conve# pol%tope over which the objective function is to be optimi4ed. K'n this conte#t, two vectors are comparable when ever% entr% in one is less*than or e!ual*to the correspondin$ entr% in the other. Otherwise, the% are incomparable.& 3inear pro$rammin$ can be applied to various fields of stud%. 't is used most e#tensivel% in business and economics, but can also be utili4ed for some en$ineerin$ problems. 'ndustries that use linear pro$rammin$ models include transportation, ener$%, telecommunications, and manufacturin$. 't has proved useful in modelin$ diverse t%pes of problems in plannin$, routin$, schedulin$, assi$nment, and desi$n. %#ality+ 7ver% linear pro$rammin$ problem, referred to as a primal problem, can be converted into a dual problem, which provides an upper bound to the optimal value of the primal problem. 'n matri# form, we can e#press the primal problem as: -a#imi4e cTx subject to x G ", x H <I with the correspondin$ sy -inimi4e "Ty subject to
T

etric dual problem, y H c, y H <.

An alternative primal formulation is: -a#imi4e cTx subject to x G "I with the correspondin$ asy -inimi4e "Ty subject to
T

etric dual problem, y C c, y H <.

There are two ideas fundamental to dualit% theor%. One is the fact that Kfor the s%mmetric dual& the dual of a dual linear pro$ram is the ori$inal primal linear pro$ram. Additionall%, ever% feasible solution for a linear pro$ram $ives a bound on the optimal value of the objective function of its dual. The weak dualit% theorem states that the objective function value of the dual

at an% feasible solution is alwa%s $reater than or e!ual to the objective function value of the primal at an% feasible solution. The stron$ dualit% theorem states that if the primal has an optimal solution, xL, then the dual also has an optimal solution, yL, such that cTxLC"TyL. A linear pro$ram can also be unbounded or infeasible. 2ualit% theor% tells us that if the primal is unbounded then the dual is infeasible b% the weak dualit% theorem. 3ikewise, if the dual is unbounded, then the primal must be infeasible. 0owever, it is possible for both the dual and the primal to be infeasible

Set-0
Q1. 8hat are the essential characteristics o& Operation Research7 Mention di&&erent phases in an Operation Research st#dy. -oint o#t so e li itations o& O.R7

(ns.

4haracteristics o& Operations Research

Operations research, an interdisciplinar% division of mathematics and science, uses statistics, al$orithms and mathematical modelin$ techni!ues to solve comple# problems for the best possible solutions. This science is basicall% concerned with optimi4in$ ma#ima and minima of the objective functions involved. 7#amples of ma#ima could be profit, performance and %ield. -inima could be loss and risk. The mana$ement of various companies has benefited immensel% from operations research. Operations research is also known as OR. 't has basic characteristics such as s%stems orientation, usin$ interdisciplinar% $roups, appl%in$ scientific methodolo$%, providin$ !uantitative answers, revelation of newer problems and the consideration of human factors in relation to the state under which research is bein$ conducted.

Syste s Orientation
o This approach reco$ni4es the fact that the behavior of an% part of the s%stem has an effect on the s%stem as a whole. This stresses the idea that the interaction between parts of the s%stem is what determines the functionin$ of the s%stem. )o sin$le part of the s%stem can have a bearin$ effect on the whole. OR attempts appraise the effect the chan$es of an% sin$le part would have on the performance of the s%stem as a whole. 't then searches for the causes of the problem that has arisen either in one part of the s%stem or in the interrelation parts.

,nterdisciplinary $ro#ps
o The team performin$ the operational research is drawn from different disciplines. The disciplines could include mathematics, ps%cholo$%, statistics, ph%sics, economics and en$ineerin$. The knowled$e of all the people involved aids the research and preparation of the scientific model.

(pplication o& Scienti&ic Methodolo$y

o OR e#tensivel% uses scientific means and methods to solve problems. -ost OR studies cannot be conducted in laboratories, and the findin$s cannot be applied to natural environments. Therefore, scientific and mathematical models are used for studies. 5imulation of these models is carried out, and the findin$s are then studied with respect to the real environment.

5ew -ro"le s Re!ealed

o (indin$ a solution to a problem in OR uncovers additional problems. To obtain ma#imum benefits from the stud%, on$oin$ and continuous research is necessar%. )ew problems must be pursued immediatel% to be resolved. A compan% lookin$ to reduce costs in manufacturin$ mi$ht discover in the process that it needs to bu% one more component to manufacture the end product. 5uch a scenario would result in une#pected costs and bud$et overruns. 7nsurin$ fle#ibilit% for such contin$encies is a ke% characteristic of OR.

-ro!ides Q#antitati!e (nswers

o The solutions found b% usin$ operations research are alwa%s !uantitative. OR considers two or more options and emphasi4es the best one. The compan% must decide which option is the best alternative for it.

.# an 2actors
o 'n other forms of !uantitative research, human factors are not considered, but in OR, human factors are a prime consideration. 6eople involved in the process ma% become sick, which would affect the compan%,s output. PHASES OPERATIONS RESEARCH M 2or #late the pro"le + This is the most important process, it is $enerall% len$th% and time consumin$. The activities that constitute this step are visits, observations, research, etc. .ith the help of such activities, the O.R. scientist $ets sufficient information and support to proceed and is better prepared to formulate the problem. This process starts with understandin$ of the or$ani4ational climate, its objectives and e#pectations. (urther, the alternative courses of action are discovered in this step.

%e!elop a odel+ Once a problem is formulated, the ne#t step is to e#press the problem into a mathematical model that represents s%stems, processes or environment in the form of e!uations, relationships or formulas. .e have to identif% both the static and d%namic structural elements, and device mathematical formulas to represent the interrelationships amon$ elements. The proposed model ma% be field tested and modified in order to work under stated environmental constraints. A model ma% also be modified if the mana$ement is not satisfied with the answer that it $ives.

Select appropriate data inp#t+ ?arba$e in and $arba$e out is a famous sa%in$. )o model will work appropriatel% if data input is not appropriate. The purpose of this step is to have sufficient input to operate and test the model. Sol#tion o& the odel+ After selectin$ the appropriate data input, the ne#t step is to find a solution. 'f the model is not behavin$ properl%, then updatin$ and modification is considered at this sta$e. Aalidation o& the odel+ A model is said to be valid if it can provide a reliable prediction of the s%stem,s performance. A model must be applicable for a lon$er time and can be updated from time to time takin$ into consideration the past, present and future aspects of the problem. , ple ent the sol#tion+ The implementation of the solution involves so man% behavioural issues and the implementin$ authorit% is responsible for resolvin$ these issues. The $ap between one who provides a solution and one who wishes to use it should be eliminated. To achieve this, O.R. scientist as well as mana$ement should pla% a positive role. A properl% implemented solution obtained throu$h O.R. techni!ues results in improved workin$ and wins the mana$ement support. 3i itations

%ependence on an Electronic 4o p#ter+ O.R. techni!ues tr% to find out an optimal solution takin$ into account all the factors. 'n the modern societ%, these factors are enormous and e#pressin$ them in !uantit% and establishin$ relationships amon$ these re!uire voluminous calculations that can onl% be handled b% computers. 5on-Q#anti&ia"le 2actors+ O.R. techni!ues provide a solution onl% when all the elements related to a problem can be !uantified. All relevant variables do not lend themselves to !uantification. (actors that cannot be !uantified find no place in O.R. models. %istance "etween Mana$er and Operations Researcher+ O.R. bein$ specialist,s job re!uires a mathematician or a statistician, who mi$ht not be aware of the business problems. 5imilarl%, a mana$er fails to understand the comple# workin$ of O.R. Thus, there is a $ap between the two. Money and /i e 4osts+ .hen the basic data are subjected to fre!uent chan$es, incorporatin$ them into the O.R. models is a costl% affair. -oreover, a fairl% $ood solution at present ma% be more desirable than a perfect O.R. solution available after sometime. , ple entation+ 'mplementation of decisions is a delicate task. 't must take into account the comple#ities of human relations and behaviour.

Q0. 8hat are the co on ethods to o"tain an initial "asic &easi"le sol#tion &or a transportation pro"le whose cost and re1#ire ent ta"le is $i!en7 Bi!e a stepwise proced#re &or one o& the 7

Ans. /ransportation -ro"le C its "asic ass# ption This model studies the minimi4ation of the cost of transportin$ a commodit% from a number of sources to several destinations. The suppl% at each source and the demand at each destination are known. The transportation problem involves m sources, each of which has available. i Ki C 8, ;, N..,m& units of homo$eneous product and n destinations, each of which re!uires bj Kj C 8, ;N., n& units of products. 0ere a i and bj are positive inte$ers. The cost cij of transportin$ one unit of the product from the ith source to the jth destination is $iven for each i and j . The objective is to develop an inte$ral transportation schedule that meets all demands from the inventor% at a minimum total transportation cost.'t is assumed that the total suppl% and the total demand are e!ual.i.e. Condition K8&The condition K8& is $uaranteed b% creatin$ either a fictitious destination with a demand e!ual to the surplus if total demand is less than the total suppl% or a Kdumm%& source with a suppl% e!ual to the shorta$e if total demand e#ceeds total suppl%. The cost of transportation from the fictitious destination to all sources and from all destinations to the fictitious sources are assumed to be 4ero so that total cost of transportation will remain the same. 2or #lation o& /ransportation -ro"le The standard mathematical model for the transportation problem is as follows. 3et #ij be number of units of the homo$enous product to be transported from source i to the destination j Then objective is to Theorem: A necessar% and sufficient condition for the e#istence of a feasible solution to the transportation problem K;& is that Q6. a. 8hat are the properties o& a $a e7 Explain the )"est strate$y* on the "asis o& in ax criterion o& opti ality. ". State the ass# ptions #nderlyin$ $a e theory. %isc#ss its i portance to "#siness decisions. (ns. aD

Mini ax Ksometimes in ax& is a decision rule used in decision theor%, $ame theor%, statistics and philosoph% for minimi4in$ the possible loss while ma"imi4in$ the potential $ain. Alternativel%, it can be thou$ht of as ma#imi4in$ the minimum $ain K axi in&. Ori$inall% formulated for two*pla%er 4ero*sum $ame theor%, coverin$ both the cases where pla%ers take alternate moves and those where the% make simultaneous moves, it has also been e#tended to more comple# $ames and to $eneral decision makin$ in the presence of uncertaint%.

Ba e theory
'n the theor% of simultaneous $ames, a minima# strate$% is a mi#ed strate$% which is part of the solution to a 4ero*sum $ame. 'n 4ero*sum $ames, the minima# solution is the same as the )ash e!uilibrium.

Mini ax theore
The minima# theorem states: (or ever% two*person, 4ero*sum $ame with finitel% man% strate$ies, there e#ists a value O and a mi#ed strate$% for each pla%er, such that Ka& ?iven pla%er ;Ps strate$%, the best pa%off possible for pla%er 8 is O, and Kb& ?iven pla%er 8Ps strate$%, the best pa%off possible for pla%er ; is QO. 7!uivalentl%, 6la%er 8Ps strate$% $uarantees him a pa%off of O re$ardless of 6la%er ;Ps strate$%, and similarl% 6la%er ; can $uarantee himself a pa%off of QO. The name minima# arises because each pla%er minimi4es the ma#imum pa%off possible for the otherRsince the $ame is 4ero*sum, he also ma#imi4es his own minimum pa%off. This theorem was established b% @ohn von )eumann,S8T who is !uoted as sa%in$ As far as ' can see, there could be no theor% of $ames N without that theorem N ' thou$ht there was nothin$ worth publishin$ until the #inima" Theorem was proved".S;T 5ee 5ion,s minima# theorem and 6arthasarath%,s theorem for $enerali4ationsI see also e#ample of a $ame without a value.

Exa ple
The followin$ e#ample of a 4ero*sum $ame, where ( and B make simultaneous moves, illustrates minima" solutions. 5uppose each pla%er has three choices and consider the pa%off matri# for ( displa%ed at ri$ht. B chooses B1 B chooses B0 B chooses B6 Assume the pa%off matri# for B is the ( chooses (1 FB Q; F; same matri# with the si$ns reversed ( chooses (0 Q8 < F> Ki.e. if the choices are A8 and /8 then ( chooses (6 Q> QB F8 B pa%s B to (&. Then, the minima# choice for ( is A; since the worst possible result is then havin$ to pa% 8, while the simple minima# choice for B is /; since the worst possible result is then no pa%ment. 0owever, this solution is not stable, since if B believes ( will choose A; then B will choose /8 to $ain 8I then if ( believes B will choose /8 then ( will choose A8 to $ain BI and then B will choose /;I and

eventuall% both pla%ers will reali4e the difficult% of makin$ a choice. 5o a more stable strate$% is needed. 5ome choices are dominated b% others and can be eliminated: ( will not choose AB since either A8 or A; will produce a better result, no matter what B choosesI B will not choose /B since some mi#tures of /8 and /; will produce a better result, no matter what ( chooses. ( can avoid havin$ to make an e#pected pa%ment of more than 8UB b% choosin$ A8 with probabilit% 8UA and A; with probabilit% VUA, no matter what B chooses. B can ensure an e#pected $ain of at least 8UB b% usin$ a randomi4ed strate$% of choosin$ /8 with probabilit% 8UB and /; with probabilit% ;UB, no matter what ( chooses. These mi#ed minima# strate$ies are now stable and cannot be improved. "D /randenbur$er and )alebuff discuss how $ame theor% works and how companies can use the principles to make decisions. The authors state that mana$ers can use the principles to create new strate$ies for competin$ where the chances for success are much hi$her than the% would be if the% continued to compete under the same rules. A classic e#ample used in the article is the case of ?eneral -otors. The automobile industr% was facin$ man% e#penses due to the incentives that were bein$ used at the retailers. ?eneral -otors responded b% issuin$ a new credit card where the cardholders could appl% a portion of their char$es towards purchasin$ a ?- car. ?- even went so far as to allow cardholders to use a smaller portion of their char$es towards purchasin$ a (ord car, allowin$ both companies to be able to raise their prices and increase lon$ term profits. This action b% ?- created a new s%stem where both ?- and (ord could be better off, unlike the traditional competitive model where one compan% must profit at the e#pense of another. The authors state that while the traditional win*lose strate$% ma% sometimes be appropriate, but that the win*win s%stem can be ideal in man% circumstances. One advanta$e to win*win strate$ies is that since the% have not been used much, the% can %ield man% previousl% unidentified opportunities. Another major advanta$e is that since other companies have the opportunit% to come out ahead as well, the% are less likel% to show resistance. The last advanta$e is that when other companies imitate the move the initial compan% benefits as well, in contrast to the initial compan% losin$ $round as the% would in a win*lose situation. The authors also state that there are five elements to competition that can be chan$ed to provide a more optimal outcome. These elements are: the pla%ers Kor companies competin$&, added values brou$ht b% each competitor, the rules under which competition takes place, the tactics used, and the scope or boundaries that are established. /% understandin$ these factors, companies can appl% different strate$ies to increase their own odds of success. The first wa% that companies can increase their chances of success involves chan$in$ who the companies are that are involved in the business. One wa% that companies can improve their odds of success is b% introducin$ new companies into the business. (or e#ample, both Coke and 6epsi wanted to $et a contract to have -onsanto as a supplier. 5ince -onsanto had a monopol% at the time, the% encoura$ed 0olland 5weetener Compan% to compete with -onsanto. 5ince it seemed -onsanto no lon$er had a monopol% on the market, the% were able to $et more favorable contracts with -onsanto. Another wa% that companies can improve their chances is b% helpin$

other companies introduce more or better complimentar% products. Companies can also chan$e the added values of themselves or their competitors. Obviousl%, companies can build a better brand or chan$e their business practices so the% operate more efficientl%. 0owever, the authors discuss how the% can also lower the value of reducin$ the value of other companies as a viable strate$%. )intendo reduced the added value of retailers b% not fillin$ all of their orders, thus leavin$ a shorta$e and reducin$ the bar$ainin$ power of the stores bu%in$ its products. The% also limited the number of licenses available to aspirin$ pro$rammers, lowerin$ their added value. The% even lowered the value held b% comic book characters when the% developed characters of their own that became widel% popular, presumabl% so that the% wouldn,t have to pa% as much to license these characters. Chan$in$ the rules is another wa% in which companies can benefit. The authors introduce the idea of judo economics, where a lar$e compan% ma% be willin$ to allow a smaller compan% to capture a small market share rather than compete b% lowerin$ its prices. As lon$ as it does not become too powerful or $reed%, a small compan% can often participate in the same market without havin$ to compete with lar$er companies on unfavorable terms. Wiwi 'nternational Air 3ines introduced services on its carriers that were of lower prices to $et market share, but made sure that the competitors understood that the% had no intention of capturin$ more than 8<X of an% market. Companies can also chan$e perceptions to make themselves better off. This can be accomplished either b% makin$ thin$s clearer or more uncertain. 'n 8YY>, the )ew York 6ost attempted to make radical price chan$es in order to $et the 2ail% )ews to raise its price to re$ain subscribers. 0owever, the 2ail% )ews misunderstood and both newspapers were headed for a price war. The )ew York 6ost had to make its intentions clear, and both papers were able to raise their prices and not lose revenue. The authors also show an e#ample of how investment banks can maintain ambi$uit% to benefit themselves. 'f the client is more optimistic than the investment bank, the bank can tr% to char$e a hi$her commission as lon$ as the client does not develop a more realistic appraisal of the compan%,s value. (inall%, companies can chan$e the boundaries within which the% compete. (or e#ample, when 5e$a was unable to $ain market share from )intendo,s :*bit s%stems, it chan$ed the $ame b% introducin$ a new 8A*bit s%stem. 't took )intendo ; %ears to respond with its own 8A*bit s%stem, which $ave 5e$a the opportunit% to capture market share and build a stron$ brand ima$e. This e#ample shows how companies can think outside the bo# to chan$e the wa% competition takes place in their industr%. /randenbur$er and )alebuff have illustrated how companies that reco$ni4e the% can chan$e the rules of competition can vastl% improve their odds of success, and sometimes respond in a wa% that benefits both themselves and the competition. 'f companies are able to develop a s%stem where the% can make both themselves and their competitors better off, then the% do not have to worr% so much about their competitors tr%in$ to counter their moves. Also, because companies can easil% cop% each other,s ideas, it is to a firm,s advanta$e if the% can benefit when their competitors cop% their idea, which is not usuall% possible under the traditional win*lose structure. This article has some parallels with the article Competin$ on Anal%tics" b% K&. The bi$$est factor that both of these articles have in common is how crucial it is for mana$ers to understand ever%thin$ the% can about their business and the environment in which the% work. 'n Competin$ on Anal%tics", the authors sa% that it is important to be familiar with this information so that mana$ers can chan$e the wa% the% compete to improve their chances of

success. At the end of The Ri$ht ?ame: Zse ?ame Theor% to 5hape 5trate$%", the authors discuss how in order for companies to be able to chan$e the environment or rules under which the% compete the% need to understand ever%thin$ the% can about the constructs under which the% are competin$. .hether a mana$er intends to use anal%tics or $ame theor% to be successful, he or she must first have all available information and use that information to understand how to make the compan% better off. 0owever, the work shown in Competin$ on Anal%tics" tends to place an emphasis almost e#clusivel% on the use of !uantitative data to improve efficienc% or market share of the compan%. The Ri$ht ?ame", however focuses more on usin$ information to find creative wa%s of chan$in$ the constructs or rules applied between companies, often %ieldin$ a much broader impact. Q4. a. 4o pare 4-M and -ER/ explainin$ si ilarities and di&&er. (ns. entionin$ where they ainly

/he Ma@or %i&&erences and Si ilarities "etween 4-M and -ER/

C6- KCritical 6ath -ethod& 1 67RTK6ro$ram 7valuation and Review Techni!ue& 8&67RT is a probabilistic tool used with three 8&C6- is a deterministic tool, with onl% sin$le 7stimatin$ the duration for completion of estimate of duration. ;&This tool is basicall% a tool for plannin$ ;&C6- also allows and e#plicit estimate of and control of time. costs in addition to time, therefore C6- can control both time and cost. B&67RT is more suitable for R12 related B&C6- is best suited for routine and those projects where the project is performed for projects where time and cost estimates can the first time and the estimate of duration be accuratel% calculated are uncertain. >&The probabilit% factor i major in 67RT >&The deterministic factor is more so values or so outcomes ma% not be e#act. outcomes are $enerall% accurate and realistic. 7#tensions of both 67RT and C6- allow the user to mana$e other resources in addition to time and mone%, to trade off resources, to anal%4e different t%pes of schedules, and to balance the use of resources. Tensions of both 67RT and C6- allow the user to mana$e other resources in addition to time and mone%, to trade off resources, to anal%4e different t%pes of schedules, and to balance the use of resources. Braphs [ 'n mathematics, networks are called $raphs, the entities are nodes, and the links are ed$es [ ?raph theor% starts in the 8:th centur%, with 3eonhard 7uler

[ The problem of W\ni$sber$ brid$es [ 5ince then $raphs have been studied e#tensivel%.

Braph /heory [ ?raph ?CKO,7& [ O C set of vertices [ 7 C set of ed$es [ An ed$e is defined b% the two vertices which it connects [ optionall%: A direction andUor a wei$ht [ Two vertices are adjacent if the% are connected b% an ed$e [ A verte#,s de$ree is the number of its ed$es Braph BEFA>ED O C set of vertices 7 C set of ed$es 7ach ed$e is now an arrow, not just a line *] direction 8 B ; > V 8 B ;

The inde$ree of a verte# is the number of incomin$ ed$es The outde$ree of a verte# is the number of out$oin$ ed$es V >