This preview shows page 1. Sign up to view the full content.
Unformatted text preview: USPS Highway Optimization
E10 IEOR Module OVERVIEW Highway corridor analytic program (HCAP) analytical model that aids USPS transportation analysts in identifying costsavings opportunities Incorporates graphical user interface Solves vehicle routingproblems with pickup and delivery Minimizes Cost model has already saved $5 million dollars in annual cost Background
USPS one of largest and most complex logistics networks in the world 200 billion pieces of mail each year Problem
Different Types of mail (letters, flats, etc.) Different Characteristics determine processing requirements Different Mail classes (1st class, 2/3 day,etc.) Different service standards for each mail class Different Transportation networks (BMC, STC, P&DC, etc.) Big Picture: Very Complicated! Objectives
Help USPS identify opportunities to reduce surfacetransportation costs Flexible in order for USPS to apply model to wide range of transportation networks Consider new transportation options Solution Approach ABC method A: Adjustables Decision Variables Facilities where mail is delivered and received Volume of mail delivered per unit time Trips/Routes taken for delivery Note: Each trip consists of series of legs. Direct trips have one leg, while multistop trips have >1 leg. ABC Method (cont'd) B Objective Function: Minimize Cost Solutions Time: Must be w/i 1% optimal Optimality: There shouldn't be any obvious improvements that can be made after model is complete Expandability: applicable for large number of trips Flexibility: Can accommodate for different constraints if introduced Availability: USPS demanded this model w/i 4 months of assignment ABC Approach (cont'd) C: Constraints Assignment Constraints: Delivery must reach the destination facility within its specified time window. Capacity constraints: For every leg of every trip the total volume associated with the deliveries routed through a trip leg must be less than or equal to the leg's capacity Solution Using CPLEX and C++ Programming Professional responsibilities: Promptness of arrival and pickup Reputation for quality service Cost and travel time minimization Ethical Responsibilities Respect for maintaining the condition of delivered goods Responsibility to conserve taxpayer's money Guaranteed arrival of essential and timesensitive packages For Future Careers
Managing constraints and limitations Increasing profitability Dealing with randomness Recognizing opportunities for flexibility Ability to balance actions and their affects on other aspects of operations Decision Traps
Plunging In Goals Transport items in a timely and cost efficient manner Meet service and financial goals Emphasis is on cost saving only, rather than on efficiency in delivery and satisfying customer needs. Priorities have not been determined Problem is defined by set of delivery requirements, facilities, and transportation resources Framing Traps
Frame Blindness/ Lack of Frame Control Influenced Directly by Decision Traps Failure to include all variables affecting cost i.e. variable weather conditions, variable traffic conditions Problem can be framed in many different terms towards cost and efficiency, may not return same results every time Which is the best way to phrase the problem, which is the most accurate way to phrase the problem? Intelligence Gathering Traps Shortsighted Shortcuts/Overconfidence in Your Judgments Standards are set based on current productivity, not on ideal productivity Emphasis is on improvement, for example: "USPS has saved millions of dollars annually" must be based on previous fiscal spending Assumption of optimality with given input data Choices to separate mail classes or combine mail classes predetermined Conclusion Traps Shooting from the Hip/Ignoring Uncertainty/ Group Failure Uncertainty in travel times Program is made to increase efficiency in travel destinations and mail order times Used for different locations, i.e. Midwest Feedback Traps Fooling Yourself About Feedback/Not Keeping Track
Based on assumption of improvement Optimal solution not defined Learning Traps Failure to Audit Your Decision Process Run several scenarios Constantly changing constraints and parameters despite limited variables taken into account to run the program Process is never completed Uncertainty
Too vast a project to take into consideration all possible factors that may alter cost and time management Weather conditions and traffic conditions not taken into account Variability in drivers effectiveness Traffic lights Division of area covered by each USPS office and affiliates Impact
Achieved higher savings Lowered costs by: Getting maximum utility, employing right amount of labor and equipment Grouping together mail orders of different delivery times together Grouping together mail orders of different weights together Finding most efficient matching legs of each delivery Flexibility of program to handle constraints Similar Problems Not unlike dilemma of travelling salesperson Some differences 1)start and finish are not same place 2) many more constraints inhibiting problem 3) no known way to find best solution, only improve upon current solution Could be used for many real world applications involving deliveries for other services Used for real world applications only, rather than idealized situations Similar to arranging airline flights to carry an unspecified number of passengers desiring to move about on a given day Package delivery like FedEx Future Application Thought processes behind HCAP Breaking down program through ABC approach Using computer programming to simply complex problems Taught us how to examine, interpret, analyze, and solve such intricate problems Works Cited
Pajunas, A., Matto, E. J., Trick, M., & Zuluaga, L. F. (2007). Optimizing Highway Transportation at the United States Postal Service. Interfaces , 37 (6), 515525. United States Postal Service Highway Corridor Analytic Program. (n.d.). IBM . ...
View Full Document
This note was uploaded on 04/03/2008 for the course E 10 taught by Professor Righter during the Spring '08 term at University of California, Berkeley.
- Spring '08