Figure 5.17 shows that if Land Shark increases its bid amount to $1,300,000, its estimatedprobability of winning the auction increases to 1.00 and its expected return increases to $89,000.Increasing the number of trials per simulation reduces the error of the estimates of theoutput. Unless the simulation model is extremely complex, it is recommended to use 10,000 trials(the maximum allowed in the educational version of ASP).To get an idea of how much noise isin the output statistics, rerun the simulation, and observe how much outputFIGURE 5.15CHART WIZARD DIALOG BOX AFTER RUNNING LAND SHARK SIMULATIONWITH ANALYTIC SOLVER PLATFORM

131statistics change. For example, in the Land Shark problem, the probability that Land Shark winsthe auction when bidding $1,250,000 is consistently between 0.4900 and 0.5050, whereas theexpected return is consistently between $77,500 and $78,500, suggesting that the 10,000 trialsis sufficient for obtaining a precise estimate.A complete version of the simulation model for the Land Shark problem can be found inthe fileLandsharkModel.The Zappos ProblemZappos is a retailer of women’s clothing, accessories, and beauty products. Suppose that theregional purchasing manager for Zappos is in the process of determining the order quantities fora new product line of women’s sleepwear for the upcoming holiday season. The new product lineconsists of four types of pajamas: cotton, flannel, silk, and velour. An initial order quantity of10,000 of each type of pajama has been proposed, but we want to evaluate this decision with asimulation model.FIGURE 5.16LAND SHARK SIMULATION OUTPUT

132The decision of how much of each pajama type to purchase depends on the pricing and demandfor each. For each pajama type, Table 5.3 lists the wholesale price (the price at which Zapposprocures the pajama from its supplier), the retail price (the price at which Zappos sells the pajamaduring the holiday season), and the clearance price (the price at which Zappos liquidates theinventory remaining after the holiday season). To help estimate the demand for each pajama type,Zappos’ marketing department has gathered a representative sample of past sales for productssimilar to the cotton and flannel pajama types. Representative data for the silk and velour pajamasare not available, so we must work with the marketing managers to gain insight on the demandfor these products.Spreadsheet Model for ZapposOur first step in modeling Zappos’ ordering decision is to identify the input parameters and outputmeasures. The next step is to develop a spreadsheet model that conveys the logical relationshipsbetween the input parameters and the output measures. Then, we prepare the spreadsheet modelfor simulation analysis by replacing the static values of the input parameters that Zappos doesnot know with certainty with probability distributions of possible values. The relevant inputparameters for Zappos are the wholesale price, retail price, clearance price, order quantity, andthe retail demand for each of the four pajama types. Total profitFIGURE 5.17OBSERVING THE IMPACT OF CHANGING BID AMOUNT WITH INTERACTIVESIMULATION

Upload your study docs or become a

Course Hero member to access this document

Upload your study docs or become a

Course Hero member to access this document

End of preview. Want to read all 202 pages?

Upload your study docs or become a

Course Hero member to access this document

Term

Summer

Professor

N/A

Tags