Question

how sensitivity analysis apply in real life? real life situation or real life scenario?

Answer 1

A iisensitivity iianalysis iidetermines iihow iidifferent iivalues iiof iian iiindependent iivariable iiaffect iia iiparticular iidependent iivariable iiunder iia iigiven iiset iiof iiassumptions. iiIn iiother iiwords, iisensitivity iianalyses iistudy iihow iivarious iisources iiof iiuncertainty iiin iia iimathematical iimodel iicontribute iito iithe iimodel's iioverall iiuncertainty. iiThis iitechnique iiis iiused iiwithin iispecific iiboundaries iithat iidepend iion iione iior iimore iiinput iivariables.

Sensitivity iianalysis iiis iiused iiin iithe iibusiness iiworld iiand iiin iithe iifield iiof iieconomics. iiIt iiis iicommonly iiused iiby iifinancial iianalysts iiand iieconomists, iiand iiis iialso iiknown iias iia iiwhat-if iianalysis.

Sensitivity iianalysis iiis iia iifinancial iimodel iithat iidetermines iihow iitarget iivariables iiare iiaffected iibased iion iichanges iiin iiother iivariables iiknown iias iiinput iivariables. iiThis iimodel iiis iialso iireferred iito iias iiwhat-if iior iisimulation iianalysis. iiIt iiis iia iiway iito iipredict iithe iioutcome iiof iia iidecision iigiven iia iicertain iirange iiof iivariables. iiBy iicreating iia iigiven iiset iiof iivariables, iian iianalyst iican iidetermine iihow iichanges iiin iione iivariable iiaffect iithe iioutcome.Assume iiSue iiis iia iisales iimanager iiwho iiwants iito iiunderstand iithe iiimpact iiof iicustomer iitraffic iion iitotal iisales. iiShe iidetermines iithat iisales iiare iia iifunction iiof iiprice iiand iitransaction iivolume. iiThe iiprice iiof iia iiwidget iiis ii$1,000, iiand iiSue iisold ii100 iilast iiyear iifor iitotal iisales iiof ii$100,000. iiSue iialso iidetermines iithat iia ii10% iiincrease iiin iicustomer iitraffic iiincreases iitransaction iivolume iiby ii5%. iiThis iiallows iiher iito iibuild iia iifinancial iimodel iiand iisensitivity iianalysis iiaround iithis iiequation iibased iion iiwhat-if iistatements. iiIt iican iitell iiher iiwhat iihappens iito iisales iiif iicustomer iitraffic iiincreases iiby ii10%, ii50%, iior ii100%. iiBased iion ii100 iitransactions iitoday, iia ii10%, ii50%, iior ii100% iiincrease iiin iicustomer iitraffic iiequates iito iian iiincrease iiin iitransactions iiby ii5%, ii25%, iior ii50% iirespectively. iiThe iisensitivity iianalysis iidemonstrates iithat iisales iiare iihighly iisensitive iito iichanges iiin iicustomer iitraffic.

performing iisensitivity iianalysis iiof iithe iisolution iiof iiLinear iiProgramming iiproblem ii(calculating iishadow iiprices, iireduced iicosts iiand iiintervals iiwithin iiwhich iithe iibasic iisolution iiremains iivalid). iiIt iiis iiclearly iidescribed iion iiacademical iiproblems iiwith ii2 iior ii3 iivariables, iibut iiin iifact, iiwhen iitried iito iiapply iithe iisame iilogic iifor iireal-life, iiscalable iiproblem, iiI iididn't iiget iipromising iiresults. iiThis iiis iibecause iionly iia iifew iiof iivariables iivalues iimatters iifor iime, iiwhile iiother iiare iirather iiplaced iifor iianother iipurposes ii(like iichanging iihard iiconstraints iito iisoft iiones iietc). iiBut iiall iiof iithem iiare iitaken iiinto iiaccount iiwhen iichecking iiif iibasic iisolution iihas iichanged, iihence iithe iiinterval iithat iiis iireturned iiby iia iisolver iiis iia iiway iimore iinarrow iithan iiI iiwant iiit iito iibe.

Sensitivity iianalysis iiis iifundamentally iiconcerned iiwith iihow iifluctuations iiin iithe iiinput iito iia iifunction iiaffect iithe iioutput iiof iithat iifunction. iiAlthough iisensitivity iianalysis iiis iiconcerned iiwith iiuncertainties iiin iiinputs iiand iioutputs iiof iia iifunction, iithe iifunction’s iibehavior iiwith iiregards iito iiinputs iiis iiresponsible iifor iithe iiuncertainties iiin iithe iioutputs, iiunless iithe iifunction iiitself iiintroduces iirandom iifluctuations, iiwhich iimight iihappen iiwhen iimodeling iia iistochastic iiprocess.

As iian iiexample iiof iia iideterministic iiprocess, iisuppose iiwe iiwant iito iiknow iithe iifloor iiarea iiof iian iiapartment iiwe iiare iiconsidering iirenting. iiIn iithe iisimplest iicase, iithe iiapartment iimight iijust iibe iia iirectangle, iiin iiwhich iicase iithere iiare iitwo iimeasurements iiwhich iiare iineeded iito iidetermine iithe iifloor iiarea: iilength iiand iiwidth. iiThe iiuncertainties iiin iithese iimeasurements iiwill iiaffect iithe iiuncertainty iiin iithe iifloor iiarea, iiwhich iimeans iiwe iicould iipotentially iiapply iisensitivity iianalysis iito iithe iiproblem iiof iimeasuring iifloor iiarea iito iidetermine iiwhich iimeasurement iicauses iithe iimost iichange iiin iithe iiestimate iiof iifloor iiarea. iiIn iithis iivery iisimple iiexample, iithe iianswer iidepends iion iithe iiroom, iisince iiabsolute iifluctuations iiin iithe iismaller iidimension iicause iilarger iichanges iiin iithe iiarea iiestimate iithan iifluctuations iiin iithe iilarger iidimension. ii(To iisee iithis, iiimagine iiadding iia iirectangle iiof iieither iiLxDW iior iiWxDL iiwhere iiL iiis iilength, iiW iiis iiwidth, iiDL iiis iia iismall iichange iiin iilength, iiand iiDW iiis iia iismall iichange iiin iiwidth. iiIf iiDL ii= iiDW, iithen iithe iilarger iidifference iiin iiarea iiis iicaused iiby iieither iiL iior iiW iibeing iilarger.)

We iicould iimake iithis iifar iimore iicomplex iihowever. iiSuppose iithat iiwe iiare iinot iimeasuring iithe iiarea iiof iian iiapartment, iibut iiinstead iiof iia iigrid iiof iiatoms iiwhich iiwe iihave iiattached iito iisome iisubstrait iito iiuse iifor iielectronics, iiexperimentation, iietc. iiWe iimight iimeasure iithis iiarea iiin iimany iiways. iiPerhaps iiwe iiuse iiScanning iiGate iiMicroscopy iito iicount iithe iiatoms, iiand iithereby iiestimate iithe iiarea iiby iicounting iiuntil iiwe iidon’t iidetect iiany iimore iiatoms iialong iia iigiven iidimension. iiThis iiprocedure iiwill iihave iimany iivariables iiwhich iicould iiaffect iithe iiestimate iiof iithe iiarea. iiIn iiall iilikelihood, iia iiMonte iiCarlo iimethod iicombined iiwith iia iimodel iifor iithe iispecific iimicroscope iimight iibe iineeded iito iiestimate iithe iiuncertainty iiof iithe iioutput iigiven iiuncertainties iiin iithe iiinputs, iiand iithis iimap iican iibe iiexplored iito iireveal iithe iispecific iieffects iiof iieach iiinput iiuncertainty.

Not iisure iiif iithese iiare iireal-world iienough iiexamples, iibut iithe iiconcept iiis iivery iigeneral, iiyet iionly iiapplicable iito iisituations iiwhere iiyou iicare iiabout iithe iiuncertainty iiin iiestimates. iiIn iimost iireal-world iisituations, iithe iiuncertainties iiinvolved iiare iinegligible iior iivery iiwell iiunderstood; iihence iiwhy iithey iiare iireal-world: iiSomeone iihas iialready iiengineered iitechnology iiand iiapplications iiof iithe iiscience, iiwhich iiwas iiexplored iiusing iiall iikinds iiof iiadvanced iistatistical iitechniques iisuch iias iisensitivity iianalysis. iiScience iiis iicomplicated, iiengineers iibridge iithe iigap iibetween iithat iicomplexity iiand iiordinary iipeople iiwho iiwant iia iisimple iiand iiuseful iitool iito iihelp iithem iiwith iithe iirest iiof iitheir iilife. iiIf iithe iitool iiitself iidoes iinot iiremove iimost iicomplexity iifor iithe iiuser, iithe iitool iiis iinot iiuseful iifor iimost iipeople.