(Approved by AICTE, New Delhi, Affiliated to Anna University, Chennai & Accredited by National Board of Accreditation (NBA, New Delhi Mailam (Po), Villupuram (Dt). Pin: 604 304 DEPARTMET !" C!MP#TER APP$%CAT%!& RESOURCE MANAGEMENT TECHNIQUES ' MC()4) Part ' A *. +,at i- operation- re-ear.,/ !perations research is a st"dy of opti#i$ation techni%"es& It is applied decision theory& !' is the application of scientific #ethods, techni%"es and tools to proble#s involvin( the operations of syste#s so as to provide these in control of operations with opti#"# sol"tions to the proble#& ). $i-t -ome appli.ation- o0 !R. !pti#al assi(n#ent of vario"s )obs to different #achines and different operators& To find the waitin( ti#e and n"#ber of c"sto#ers waitin( in the %"e"e and syste# in %"e"in( #odel To find the #i#i#"# transportation cost after allocatin( (oods fro# different ori(ins to vario"s destinations in transportation #odel Decision theory proble#s in #ar*etin(,finance and prod"ction plannin( and control& 3. +,at are t,e 1ariou- t2pe- o0 mo3el- in !R/ +odels by f"nction i Descriptive #odel ii ,redictive #odel iii Nor#ative #odel +odels by str"ct"re i Iconic #odel ii Analo("e #odel iii +athe#atical #odel +odels by nat"re of environ#ent i Deter#inistic #odel ii ,robabilistic #odel 4. +,at are main .,ara.teri-ti.- o0 !R/ E-a#ination of f"nctional relationship fro# a syste# overview& Utili$ation of planned approach Adaptation of planned approach Uncoverin( of new proble#s for st"dy 4. ame -ome .,ara.teri-ti.- o0 goo3 mo3el. The n"#ber of ass"#ptions #ade sho"ld be as few as possible It sho"ld be easy as possible to solve the proble# The n"#ber of variables "sed sho"ld be as few as possible& Prepared By Ms.A.Subathra Devi AP / MCA UNIT I - LINEAR PROGRAMMING MODELS Mathematical Formulation - Graphical Solution of linear programming models Simplex method Artificial variable Techniques- Variants of Simplex method 2 It sho"ld be #ore fle-ible to "pdate the chan(es over a period of ti#e witho"t chan(e in its fra#ewor*& 6. +,at are t,e 3i00erent p,a-e- o0 !R/ .or#"lation of the proble# Constr"ction of #athe#atical #odelin( Derivin( the sol"tion fro# the #odel /alidity of the #odel Establishin( the control over the sol"tion I#ple#entation of the final sol"tion& 5. $i-t out t,e a31antage- o0 !R/ !pti#"# "se of #ana(ers prod"ction factors I#proved %"ality of decision ,reparation of f"t"re #ana(ers by i#provin( their *nowled(e and s*ill +odification of #athe#atical sol"tion before its "se& 6. +,at are t,e limitation- o0 !R/ +athe#atical #odel do not ta*e into acco"nt the intan(ible factors s"ch as h"#an relations etc& cannot be %"antified& +athe#atical #odels are applicable to only specific cate(ories of proble#s& 'e%"ires h"(e calc"lations& All these calc"lations cannot be handled #an"ally and re%"ire co#p"ters which bear heavy cost& (. +,at i- linear programming/ 0inear pro(ra##in( is a techni%"e "sed for deter#inin( opti#"# "tili$ation of li#ited reso"rces to #eet o"t the (iven ob)ectives& The ob)ective is to #a-i#i$e the profit or #ini#i$e the reso"rces (#en, #achine, #aterials and #oney *0. +rite t,e general mat,emati.al 0ormulation o0 $PP. 1& !b)ective f"nction +a- or +in 2 3 C1-1 4 C5-54 6&&4 Cn-n 5& 7"b)ect to the constraints a11-14a15-5466664 a1n-n (839b1 a51-14a55-5466664 a5n-n (839b5 6666666666666666666666&& 6666666666666666666666&& a#1-14a#5-5466664 a#n-n (839b# :& Non;ne(ative constraints -1,-5,6&-#9 < **. +,at are t,e .,ara.teri-ti. o0 $PP/ There #"st be a well;defined ob)ective f"nction& There #"st be alternative co"rse of action to choose& Both the ob)ective f"nctions and the constraints #"st be linear e%"ation or ine%"alities& *). +,at are t,e .,ara.teri-ti. o0 -tan3ar3 0orm o0 $PP/ The ob)ective f"nction is of #a-i#i$ation type& All the constraint e%"ation #"st be of e%"al type by addin( slac* or s"rpl"s variables '=7 of the constraint e%"ation #"st be positive type Prepared By Ms.A.Subathra Devi AP / MCA 3 All the decision variables are of positive type *3. +,at are t,e .,ara.teri-ti.- o0 .anoni.al 0orm o0 $PP/ (!V 705) In canonical for#, if the ob)ective f"nction is of #a-i#i$ation type, then all constraints are of 8 type& 7i#ilarly if the ob)ective f"nction is of #ini#i$ation type, then all constraints are of 9 type& B"t non;ne(ative constraints are 9type for both cases& *4. A 0irm manu0a.ture- t8o t2pe- o0 pro3u.t- A an3 9 an3 -ell- t,em at pro0it o0 R- ) on t2pe A an3 R- 3 on t2pe 9. Ea., pro3u.t i- pro.e--e3 on t8o ma.,ine- M* an3 M).T2pe A re:uire- * minute o0 pro.e--ing time on M* an3 ) minute- on M) T2pe 9 re:uire- * minute o0 pro.e--ing time on M* an3 * minute on M). Ma.,ine M* i- a1aila;le 0or not more t,an 6 ,our- 40 minute- 8,ile ma.,ine M) i- a1aila;le 0or *0 ,our- 3uring an2 8or<ing 3a2. "ormulate t,e pro;lem a- a $PP -o a- to ma=imi>e t,e pro0it. (MA? 705) +a-i#i$e $ 35-1 4:-5 7"b)ect tot the constraints> -1 4 -5 8 ?<< 5-1 4 -5 8 @<< -1 ,-59 < *4. A .ompan2 -ell- t8o 3i00erent pro3u.t- A an3 9 , ma<ing a pro0it o0 R-.40 an3 R-. 30 per unit on t,em,re-pe.ti1el2.T,e2 are pro3u.e3 in a .ommon pro3u.tion pro.e-- an3 are -ol3 in t8o 3i00erent mar<et-, t,e pro3u.tion pro.e-- ,a- a total .apa.it2 o0 30,000 man@,our-. %t ta<e- t,ree ,our- to pro3u.e a unit o0 A an3 one ,our to pro3u.e a unit o0 9. T,e mar<et ,a- ;een -ur1e2e3 an3 .ompan2 o00i.ial 0eel t,at t,e ma=imum num;er o0 unit- o0 A t,at .an ;e -ol3 i- 6,000 unit- an3 t,at o0 9 i- *),000 unit-. &u;Ae.t to t,e-e limitation-, pro3u.t- .an ;e -ol3 in an2 .om;ination. "ormulate t,e pro;lem a- a $PP -o a- to ma=imi>e t,e pro0it +a-i#i$e $ 3?<-1 4:<-5 7"b)ect tot the constraints> :-1 4 -5 8 :<,<<< -1 8 A<<< -5 8 15<<< -1 ,-59 < *6. +,at i- 0ea-i;ilit2 region/ (MA? 706) Collections of all feasible sol"tions are called a feasible set or re(ion of an opti#i$ation #odel& !r A re(ion in which all the constraints are satisfied is called feasible re(ion& *5. +,at i- 0ea-i;ilit2 region in an $P pro;lem/ %- ti ne.e--ar2 t,at it -,oul3 al8a2- ;e a .on1e= -et/ A re(ion in which all the constraints are satisfied is called feasible re(ion& The feasible re(ion of an 0,, is always conve- set& *6. De0ine -olution A set of variables -1,-56&-n which satisfies the constraints of 0,, is called a sol"tion& *(. De0ine 0ea-i;le -olution/ (MA? 705) Any sol"tion to a 0,, which satisfies the non;ne(ativity restrictions of 0,,Bs called the feasible sol"tion Prepared By Ms.A.Subathra Devi AP / MCA 4 )0. De0ine optimal -olution o0 $PP. (MA? 70() Any feasible sol"tion which opti#i$es the ob)ective f"nction of the 0,,Bs called the opti#al sol"tion )*. &tate t,e appli.ation- o0 linear programming Cor* sched"lin( ,rod"ction plannin( & prod"ction process Capital b"d(etin( .inancial plannin( Blendin( .ar# plannin( Distrib"tion +"lti;period decision proble# Inventory #odel .inancial #odel Cor* sched"lin( )). &tate t,e $imitation- o0 $P. 0, treats all f"nctional relations as linear 0, does not ta*e into acco"nt the effect of ti#e and "ncertainty No ("arantee for inte(er sol"tion& 'o"ndin( off #ay not feasible or opti#al sol"tion& Deals with sin(le ob)ective, while in real life the sit"ation #ay be diffic"lt& )3. +,at 3o 2ou un3er-tan3 ;2 re3un3ant .on-traint-/ In a (iven 0,, any constraint does not affect the feasible re(ion or sol"tion space then the constraint is said to be a red"ndant constraint& )4. De0ine #n;oun3e3 -olution/ If the feasible sol"tion re(ion does not have a bo"nded area the #a-i#"# val"e of 2 occ"rs at infinity& =ence the 0,, is said to have "nbo"nded sol"tion&
)4. De0ine Multiple !ptimal -olution/ A 0,, havin( #ore than one opti#al sol"tion is said to have alternative or #"ltiple opti#al sol"tions& )6& +,at i- -la.< 1aria;le/ If the constraint as (eneral 0,, be D3 type then a non ne(ative variable is introd"ced to convert the ine%"alities into e%"alities are called slac* variables& The val"es of these variables are interpretedas the a#o"nt of "n"sed reso"rces& )5. +,at are -urplu- 1aria;le-/ If the constraint as (eneral 0,, be E3 type then a non ne(ative is introd"ced to convert the ine%"alities into e%"alities are called the s"rpl"s variables&
)6. De0ine 9a-i. -olution/ Fiven a syste# of # linear e%"ations with n variables(#Dn&The sol"tion obtained by settin( (n;# variables e%"al to $ero and solvin( for the re#ainin( # variables is called a basic sol"tion& )(& De0ine non Degenerate 9a-i. 0ea-i;le -olution/ Prepared By Ms.A.Subathra Devi AP / MCA 5 The basic sol"tion is said to be a non de(enerate basic sol"tion if None of the basic variables is $ero& 30. De0ine 3egenerate ;a-i. -olution/ A basic sol"tion is said to be a de(enerate basic sol"tion if one or #ore of the basic variables are $ero& 3*. +,at i- t,e 0un.tion o0 minimum ratioG To deter#ine the basic variable to leave To deter#ine the #a-i#"# increase in basic variable To #aintain the feasibility of followin( sol"tion 3). "rom t,e optimum -imple= ta;le ,o8 3o 2ou i3enti02 t,at $PP ,a- un;oun3e3 -olution/ To find the leavin( variables the ratio is co#p"ted& The ratio is D3< then there is an "nbo"nded sol"tion to the (iven 0,,& 33. "rom t,e optimum -imple= ta;le ,o8 3o 2ou i3enti02 t,at t,e $PP ,a- no -olution/ If atleast one artificial variable appears in the basis at $ero level with a 4ve val"e in the Hb col"#n and the opti#ality condition is satisfied then the ori(inal proble# has no feasible sol"tion& 34. Bo8 3o 2ou i3enti02 t,at $PP ,a- no -olution in a t8o p,a-e met,o3/ If all 2) I C) 8 < & then atleast one artificial variable appears in the opti#"# basis at non $ero level the 0,, does not possess any sol"tion& 34& +,at 3o 2ou un3er-tan3 ;2 3egenera.2/ The concept of obtainin( a de(enerate basic feasible sol"tion in 0,, is *nown as de(eneracy& This #ay occ"r in the initial sta(e when atleast one basic variable is $ero in the initial basic feasible sol"tion& 36. +rite t,e -tan3ar3 0orm o0 $PP in t,e matri= notation/ In #atri- notation the canonical for# of 0,, can be e-pressed as +a-i#i$e 2 3 CH(ob) fn& 7"b to AH D3 b(constraints and H E3 < (non ne(ative restrictions Chere C 3 (C1,C5,6&&Cn,
A 3 a11 a15 6&& a1n H 3 -1 b 3 b1 a51 a556&& a5n , -5 , b5 & & & & & & a#1 a#56& a#n -n bn 35. De0ine ;a-i. 1aria;le an3 non@;a-i. 1aria;le in linear programming. A basic sol"tion to the set of constraints is a sol"tion obtained by settin( any n variables e%"al to $ero and solvin( for re#ainin( # variables not e%"al to $ero& 7"ch # variables are called basic variables and re#ainin( n $ero variables are called non;basic variables. 36.&ol1e t,e 0ollo8ing $P pro;lem ;2 grap,i.al met,o3. (MA? 706) Ma=imi>e > C6=* D4=) &u;Ae.t tot t,e .on-traint-: Prepared By Ms.A.Subathra Devi AP / MCA 6 =* D =) E 4 =)F 6 =* ,=)F 0 3(. De0ine unre-tri.te3 1aria;le an3 arti0i.ial 1aria;le. (!V 705) Unrestricted /ariable >A variable is "nrestricted if it is allowed to ta*e on positive, ne(ative or $ero val"es Artificial variable >!ne type of variable introd"ced in a linear pro(ra# #odel in order to find an initial basic feasible sol"tionJ an artificial variable is "sed for e%"ality constraints and for (reater;than or e%"al ine%"ality constraints 40. +,at i- t,e u-e o0 arti0i.ial 1aria;le- in $PP/ GMa2 )0**H The "se of artificial variables is to obtain an initial basic feasible sol"tion& In addition to these artificial variables ca"ses violation of the correspondin( constraints& 4 Prepared By Ms.A.Subathra Devi AP / MCA