Next Wave Agency Ocean Carrier Bid Optimization April

Next Wave Agency Ocean Carrier Bid Optimization April 16, 2008 Final Presentation Senior Design Team: Juan Araya Steven Butts Owen Carroll Emily Sarver Justin Stowe Jan Zhang Sponsor: John Trestrail, CEO & Principal Consultant jtrestrail@nextwaveagency. com Advisor: Dr. Ozlem Ergun ozlem. ergun@isye. gatech. edu *This presentation was created in the framework of a student design project. The Georgia Institute of Technology does not sanction its content in any way.

Outline Company and Problem Background Project Description Solution Approach Deliverables Value and Benefits 2

Food for Peace Program USDA & USAID Invitation published monthly to procure: Food commodities Transportation USDA awards contracts to minimize cost 3

Food for Peace Program 4

Food for Peace Program 5

Next Wave Agency Background Specializes in Food for Peace Program Ocean carrier consulting agency Recommends bid prices to carriers 6

Project Introduction Bids based on market knowledge Wanted to develop an analytical method Created a tool to determine bid prices 7

Project Flowchart Supplier Bid Forecasting User Input Optimization Model Analysis Carrier Bid Forecasting Bid Recommendations 8

Project Flowchart Supplier Bid Forecasting User Input Optimization Model Analysis Carrier Bid Forecasting Bid Recommendations 9

Solution Approach: Forecasting Parameters for optimization model Supplier bids Origin ports Commodities Carrier bids 10

Solution Approach: Forecasting Supplier Bids Data available Winning and losing bids for past year Bids are for a commodity at an origin 3, 750 pairs 10 suppliers per pair Supplier interaction 11

Solution Approach: Forecasting Supplier Bids Minimum bids for origin-commodity pairs Upward trend Non-seasonal 12

Solution Approach: Forecasting Supplier Bids Double exponential smoothing Non-seasonal data Trend in minimum supplier bids over time Low mean absolute percent error (5. 62%) 13

Solution Approach: Forecasting Carrier Bids Data available Past winning carrier bids Past market indices for ocean freight costs Voyage lengths Past voyage factors highly variable Analyze past voyage data 3 carriers 10 vessels 14

Solution Approach: Forecasting Carrier Bids Estimate costs per ton Fuel Daily leasing Port call costs Method considered: Regression High R-squared (87 -99. 7%) Not adaptable for unusual voyages 15

Solution Approach: Forecasting Carrier Bids Use market costs to estimate profit per ton Past profit per ton fits normal distribution 16

Solution Approach: Forecasting Carrier Bids Method developed Calculate voyage costs using current indices Add random variable for profit per ton Calculate confidence intervals for bid prices Carrier Maersk Sealift Liberty Expected Profit 80% CI 95% CI 128. 70 ± 22. 14 ± 28. 23 ± 33. 64 153. 80 ± 12. 08 ± 15. 41 ± 18. 36 144. 80 ± 19. 46 ± 24. 81 ± 29. 56 17

Solution Approach: Forecasting Carrier Bids Houston to Djibouti 20, 000 tons Maersk vessel Fuel Lease Port call Total costs per ton Add Expected Profit Expected Bid Price $29. 18 $44. 12 $5. 00 $78. 30 $128. 70 $207. 00 18

Solution Approach: Forecasting Carrier Bids Confidence intervals for expected bid price Expected Bid Price $207. 00 80% Confidence Interval (195, 219) 90% Confidence Interval (192, 222) 95% Confidence Interval (189, 225) 19

Project Flowchart Supplier Bid Forecasting User Input Optimization Model Analysis Carrier Bid Forecasting Bid Recommendations 20

Project Flowchart Supplier Bid Forecasting User Input Optimization Model Analysis Carrier Bid Forecasting Bid Recommendations 21

Solution Approach: Optimization Simulates USDA’s process of awarding contracts Objective: Minimize total cost Constraints: Carrier quantity Commodity demand Carrier-supplier pairing US vessel priority 22

Solution Approach: Optimization Validated using actual data from 3 past invitations Checked tonnage distribution among carriers Average accuracy for each carrier: 91. 7% 23

Solution Approach: Optimization Tested with forecasted supplier bids Change in error: <1% 24

Solution Approach: Optimization Run model multiple times Increment bid prices for Next Wave Analyze results Determine bid prices 25

Project Flowchart Supplier Bid Forecasting User Input Optimization Model Analysis Carrier Bid Forecasting Bid Recommendations 26

Project Flowchart Supplier Bid Forecasting User Input Optimization Model Analysis Carrier Bid Forecasting Bid Recommendations 27

Deliverables 28

Deliverables Expected Price 29

Deliverables 30

Deliverables 31

Deliverables 32

Deliverables 33

Deliverables 34

Deliverables 35

Value and Benefits Financial value from increase in bids won If 10% increase: $550, 000 for clients $5, 500 for Next Wave Unique advantage over competitors Potential to attract new clients 36

Summary Created analytical bidding tool Forecasting bids Optimization model Software User Interface Database Reports 37

Questions? 38