“
(16.1
Mee
16.1 Risk Analysis n5
Simulation is ot an optimization technique. Itis a method shat can be used 1 describe
or predict how a systom will operate given certain choices for the controllable inputs and
rardomly generaied values for the probebilistic inputs. Analysts offea use simulation t0
etermine values for the controllable inputs that are likely to lead to desirable system
‘outputs. In this sense, simulation ean be an effective tool in designing a system to provide
200d performance.
In tischapter we begin by showing bow simulation caa be used to sty the financial risks
ascocizted with the development ofa new product. We continue withillustraons showing how
‘Simulation can be used to estzblish an effective inventory policy and bow itcan be used t9 de-
sign waiting line systems. We conclude with a discussion of other isues, such as very
‘simulation program, validating the mode}, and selecting a simulation software package
Risk Analysis
‘Risk analysis is the process of predicting the outcome of a decision in the face of uncer-
tainty In this section we describe a problem that invelves considerable uncertainty: the de-
velopment of a new product. We first show how risk analysis can be conducted without
using simulation; we then show how a more comprehensive risk analysis can be conducied
‘with the aid of simulation.
PortaCom Project
PortaCom mianufacurres notebook computers and related equipment. PortaCom’s product
design aroup developed pretorype for new high-quality portable printer. The new printer
Features an innovative design and has the potential to capture a significant share of the
portable printer market, Preliminary marketing and financial anelyses provided the follow-
ing selling price, first-year administrative cost, and first-year advertising cost
Selling price = $249 per unit
Administrative cost = $400,000
‘Advertising cost = $600,000
In the simulation model for the PortzCom problem. the preceding values are constants and
are referred to as parameters of the model.
‘The cost of dizect labor, the cost of parts, and the first-yeer demand for the printer are
hot known with certainty and are considered probabiliscie inputs. At this stage of the plan
ning process, PortaCom's best estimates of these inputs are $45 per unit forthe direct labor
‘cost, $90 per unit for the parts cost, and 15,000 units for the first-year demand. PortaCom
‘would lke an analysis of the first-year profit potential for the printer. Because of PortaCom’s
tight cash flow situntion, management is particularly concerned about the potential for loss,
What-lf Analysis
One approach to risk analysis is called what-if analgsis. A what-if analysis involves ge
erating values for the probabilistic inpu's (direct labor cos, paris cost, and fi
demand) and competing the resulting value forthe ouput (prot), Wit a selling price Of
$249 per oni: and administrative pls advertising costs equal co $100,000 + $606,000
$1,000,000, the PortaCon: profit model is
Profit = (S249 — Direct labor cost per unit ~ Paris cost per unit)(Demand) — $1,000,00076
Cheploe 15 Simulation
FIGURE 16.2 PORTACOM PROFIT MODEL
Probabilistic
puts
Direct nt
Tato: Fars _ Year
Gest Cost Demand
{ii
Iprgtiee lp eye ~ 1,000,000} — rote
Leting
6, = direct labor cost per unit
6 = pats cow per unit
x= first-year demand
ihe profit model for ce frst year can be written as follows
Profit = (249 ~ c — eqhx ~ 1,000,000 (16.1)
‘The PoxsCom profit model ean be depicted as shown in Figure 16.2
Recall that PortaCom's best estimates of the direct labor cost per uni, the parts cost p:
‘ani, end first-year demand are S45, $90, and 15.000 units, respectively. These values eo
stitute the hase-eace scenario for PortaCom. Substituting these values into equation (16.
ylelds the following profit projection:
Profit = (249 ~ 45 ~ 90)(15,000) ~ 1.000.000 = 710,000
Thus, the buse-case scenario leads 0 au anticipated profit of $710,006
Inrisk analysis we are concemed with both the probability of loss and the magnituc
of a loss. Although the base-case scenario looks sppealing. PortuCom might be interest:
in what hapgeus if the estimates of the direst labor cost per unit, pats cost per unit, ar
rstyear demand do not tum ont t9 be as expecied uncer the buse-case scenario. F
instance, suppose that PortaCom believes that direct labor costs could range fiom $43,
S47 per unit, parts cost could range from $80 to $100 per unit, and first-year demand cou
range from 1500 ro 28,500 units. Using these ranges. what-it analysis can be used
evaluate a Worst-case scenario and a best-case scenario,
The worst-case value for the divecr labor cost is S47 ((he highest value), the worst-ea:
value forthe paris cost is $100 (the highest value), and the worst-case value for demand
1500 wits the lowest value). Thus, in the worst-case scenaio, ¢; = 47,€3 ~ 100, and.
1500. Substituting chose values into equation (6.1) leads to the following profit projestio
Profit = (249 — 47 — 190)(1500) ~ 1,000,000 = —847,000
So the worst-case scenario less to.a projected lose of $847,000,
The best-case value for the direct labor cost is $43 (the lowest value), the besteca
value for the puts cost is $80 (the lowest value). and the best-cose valve for demandProbl 2 wil give x08
practesashig whee
‘nal
One abvanese of
simalsion isthe abil
diszituvons thet are
nique the system being
161
Anolyie nr
28,500 units (the highest value). Subetitting
Following protic projection:
Profit = (219 — 43 — 804:
hase values into equation (16.1) leads to the
10) ~ 1,000,000 = 2.591,000
So the best-case scenario leads to 2 projected profit of $2,591,000,
At this point the sshati analysis provides the conclusion that profits can rar
Joss of $847,000 to a profi of S2.591.000 with a base-case profit of $710,000. Although
the base-case profit of $710.000 is possible, the what-if analysis indicates that either a
substantial loss or a substantial profit is possible. Other scenarios that PortaCom might
‘Want to consider can also be evaluated. However. the difficulty with whee if analysis is that
it does act indicate the likelihood of the various profit or oss values. In particular we de
not Enow anything about the probability ofa loss.
mulation to perform risk analysis for the PortsCom problem is like playing ovt many
enaos by randomly generating values lor the probsbilistic inputs, The advantage of
Simulation is tet it alos us to assess the probability of a profit and the pmobubilty ofa less
Using the what-if approach io risk analysis, we selected values forthe probabilistic in~
puis (direct tabor cost per unit |. parts cos! per unit (c,), and first-year demand (s)} and
then computed the resulting profit, Applying simulation to the PortaCom problem requires
generating values forthe probabilistic inputs that are representative of what we might ob-
serve in practice, To generate such values. we must know the probability distribution for
‘each probabilistic input, Further analysis by PortaCom ledto the following probability dis
butions forthe direct labor cust per unit. tke parts cost per unit, and First-year demand:
Direct Labor Cost PostaCom believes thu the direct labor cost will range from $43 to $47
Der nit and js described by the discrots probability distribution shown in Table 16.1. Thusy
we see a O11 probability thatthe direct labor cost will be $43 per unit, a 0.2 probability that
the direct labor cost will be $44 per unit, and so on, The higltest probability of 4 is asso-
ciated with a direct labor cost of S45 per unit.
Parts Cost This cost depends upon the general economy. the overall demund for pars, and
the pricing policy of PortaCom's parts suppliers. PoetaCom believes tha the parts eost will
range from $8) to $100 per unit aud iy described by the uniform probability distribution
shown in Figure 16,3. Costs per unit between $80 and $100 are equally likely.
First-Year Demand PoraCom believes tha first-year demand is desciibed by the norma
Probability distribution shown in Figure [6.4. The mean or expecied value of first-year
TABLE 16.1 PROBABILITY DISTRIBUTION FOR DIRECT LABOR COST PER UNIT
Direct Labor Cost
per Unit,
su
Sti
SS
S16
Su