IMPLEMENTING STRATEGIC PLANNING UNDER UNCERTAINTY USING PALISADE’S RISKOPTIMIZER® Tayo Fabusuyi Numeritics Palisade’s Risk Conference March 2011 Outline Introduction and Context Objectives Sources of Uncertainty Model Description Model Output Discussion of Findings Introduction and Context Manufacturing entity that produces degreasers, descalers, neutralizers and industrial cleaners Firm recently changed ownership with previous products re-branded Products are sold in 5, 15, 55 and 275 gallon containers Customers are primarily wholesale (regional or national chains) and corporate clients Objectives Positive Economic Profit Revenue Maximization Uncertain Variables Cost per unit of product Health premiums, utilities cost (dynamic pricing), labor cost (overtime), transportation/gasoline related cost, variations in raw material prices etc. Uncertainty addressed using normal distribution Demand Variability distribution in demand addressed using triangular Model Description Max: Subject to: Model Modifications Unmet Demand Structured as a soft constraint with attached penalty Profit Variability Included as a hard constraint with a specified standard deviation threshold Model Implementation Begining Inventory Produced Total Available Units Sold Sales price Raw Material 1 Raw Material 2 Labor (hours) Machine Time Cost Price Profit/unit sold Overage Unit storage cost Overage Cost Actual Demand Worst Most Likely Best Unmet Demand Descaler (5) Descaler (15) Descaler (55) Descaler (275) Degreaser (5) 34 1478 1512 1212 2.75 0.022 0.000 0.25 0.1875 2.02 0.73 300 0.12 36.00 1212 916 1280 1440 73 960 1033 862 6.15 0.066 0.000 0.32 0.2400 3.64 2.51 171 0.26 44.55 862 642 895 1048 0 1616 1616 521 13.42 0.242 0.000 0.72 0.5400 7.78 5.64 1095 0.33 361.24 521 394 550 620 112 1283 1395 773 21.70 1.210 0.000 1.32 0.9900 13.33 8.37 622 0.47 292.50 773 583 822 913 5 997 1002 622 1.71 0.000 0.018 0.25 0.1875 0.94 0.77 380 0.12 45.56 622 470 657 740 0 0 0 0 0 Model Implementation (cont.) Model Implementation (cont.) Utilized Available Raw Material 1 Raw Material 2 Labor (Man Hour) Machine Time 2415.22 1672.96 8389.67 6532.27 Cost of Capital Aggregate Unmet Demand (Boolean) Demand Shortage Penalty Threshold 6333.33 0 0 0.25 Revenue Profit Overage Cost NOPAT $ $ $ 64,744.64 19,951.05 1485.69 11,970.63 3000 2000 12800 7200 Mean (Best Simulation) = 64745.8007 Model Output Results Valid Simulations 18 Total Simulations 32 Original Value $ 52,395.75 + soft constraint penalties $ = result $ 51,645.75 $ 64,745.80 Best Value Found + soft constraint penalties $ = result $ Best Simulation Number Time to Find Best Value Reason Optimization Stopped (750.00) 64,745.80 26 0:08:44 Elapsed time Time Optimization Started 3/24/2011 16:45 Time Optimization Finished 3/24/2011 16:55 Total Optimization Time 0:10:09 Adjustable Cell Values P r oduct O r igina l Bes t Descaler (5) 1270 1478 Descaler (15) 842 960 Descaler (55) 1616 1616 Descaler (275) 804 1283 Degreaser (5) 997 997 Degreaser (15) 538 1217 Degreaser (55) 448 875 Degreaser (275) 359 622 Dual (5) 324 1375 Dual (15) 591 591 Dual (55) 912 912 Dual (275) 324 795 Resource Constraint Findings Resource Constraints Description Raw Material 1 Labor = 0 <= 'Rev'!$C$26 <= 'Rev'!$D$26 = 0 <= 'Rev'!$C$28 <= 'Rev'!$D$28 Constraint Type Hard Hard Evaluation Time Iteration Iteration 87.50% 75.00% 100.00% 100.00% Raw Material 2 Machine Time = 0 <= 'Rev'!$C$27 <= 'Rev'!$D$27 = 0 <= 'Rev'!$C$29 <= 'Rev'!$D$29 Constraint Type Hard Hard Evaluation Time Iteration Iteration 62.50% 66.67% 100.00% 100.00% Definition Satisfied for % of Simulations Satisfied for % of Valid Simulations Description Definition Satisfied for % of Simulations Satisfied for % of Valid Simulations Results for Goal Related Constraints Goal Related Constraints Description EVA Definition ='Rev'!$C$41 >= 'Rev'!$C$32 Constraint Type Hard Evaluation Time Iteration Satisfied for % of Simulations 100.00% Satisfied for % of Valid Simulations 100.00% Description NOPAT DEV Definition =RiskStdDev('Rev'!$C$41) <= 1500 Constraint Type Hard Evaluation Time Simulation Satisfied for % of Simulations 50.00% Satisfied for % of Valid Simulations 100.00% Description Shortage Penalty Definition ='Rev'!$C$34 <= 0.25 Constraint Type Soft Evaluation Time Simulation Satisfied for % of Simulations 29.17% Satisfied for % of Valid Simulations 25.00% Penalty Function Penalty of Best Result =DEVIATION*1000 $ - Facts behind the figures All production steps for which there were excess capacity or where real time changes could be quickly effected were considered to be neither restrictive nor critical and were therefore omitted Results obtained may not be implementable in every detail but should be sufficiently valid for bench-marking purposes Value Added by the approach Generated the ideal quantity to be produced of each product sold and provided insights pointing to a more productive use of existing facilities Reveal potential bottlenecks in the production process Invaluable information was revealed by relaxing the demand constraint Provision of a robust vehicle for evaluating alternative production plans Questions & Contact Information Questions? Tayo Fabusuyi Numeritics, Pittsburgh, PA Tayo.Fabusuyi@numeritics.com +1.412.874.4417