1. The Computerized Routing of Meter Readers over Street Networks: New Technologies Turn Old Problems into New Opportunities by Robert Shuttleworth, University of Maryland Bruce Golden, University of Maryland Edward Wasil, American University Presented at EURO XXII Prague, July 2007

2. 2 Outline of Lecture  The close enough traveling salesman problem (CETSP)  The CETSP over a street network  Heuristics for solving this problem  Greedy Approaches  IP Formulations  Computational Results  Conclusions

3. 3  Until recently, utility meter readers had to visit each customer location and read the meter at that site  Now, radio frequency identification (RFID) technology allows the meter reader to get close to each customer and remotely read the meter  Our models are based on data from a utility and use an actual road network with a central depot and a fixed radius r for the hand held device  Our goal is to minimize distance traveled or elapsed time The CETSP over a Street Network

4. 4  We used RouteSmart (RS) with ArcGIS  Real-world data and constraints  Address matching  Side-of-street level routing  Solved as an arc routing problem  Our heuristic selects segments to exploit the “close enough” feature of RFID  RS routes over the chosen segments to obtain a cycle  Currently, RS solves the problem as a Chinese (or rural) Postman Problem The CETSP over a Street Network

5. 5  How do we choose the street segments to feed into RS?  We tested several ideas  Greedy procedures  Greedy A: Choose the street segment that covers the most customers, remove those customers, and repeat until all customers are covered  Greedy B: Same as above, but order street segments based on the number of customers covered per unit length  IP Formulations Heuristic Implementation

6. 6  We also experimented with formulating the problem as an IP: IP Formulation

7. 7  We tested several choices for the objective function  IP1: Minimize the number of road segments chosen c j = 1 for all j  IPD1: Minimize the distance of the road segments chosen c j = the distance of road segment j IP Variants

8. 8 Each Color is a Separate Partition

9. 9 A Single Partition

10. 10 A Closer Look at a Partition

11. 11 The Area Covered with RFID

12. 12 The Area Covered by the Entire Partition

13. 13 500 foot radius MethodMilesHours Number of Segments Miles of Segments Deadhead Miles RS204.89:22109997.5107.3 Greedy A160.57:06470 64.496.1 Greedy B166.57:27577 64.2102.3 IP1165.87:25458 62.4103.4 IPD1161.67:15470 59.1102.5 Essential−−34243.3− Dense Partition Results

14. 14 Dense Partition Results 350 foot radius MethodMilesHours Number of Segments Miles of Segments Deadhead Miles RS204.89:22109997.5107.3 Greedy A171.97:4562178.193.8 Greedy B179.37:5561078.0101.3 IP1169.87:3960877.692.2 IPD1168.17:4060976.991.2 Essential−−45161.9−

15. 15 500 foot radius MethodMilesHours Number of Segments Miles of Segments Deadhead Miles RS213.69:2640598.4115.2 Greedy A189.98:2221779.6110.3 Greedy B197.08:5623684.7112.3 IP1188.28:1821678.5109.7 IPD1188.48:1821678.3110.1 Essential−−21278.0− Sparse Partition Results

16. 16 Sparse Partition Results 350 foot radius MethodMilesHours Number of Segments Miles of Segments Deadhead Miles RS213.69:2640598.4115.2 Greedy A200.18:3437991.2108.9 Greedy B203.18:4139193.3109.8 IP1200.58:3637891.6108.9 IPD1201.08:3738091.0110.0 Essential−−32585.9−

17. 17 500 foot radius MethodMilesHours Number of Segments Miles of Segments Deadhead Miles RS3798.1165:41165091545.12253.0 Greedy A3045.2140:0598951498.91546.3 Greedy B3140.3144:41114831528.61611.7 IP13055.6140:3798571492.81562.8 IPD13039.1140:0299071491.81547.3 Essential−−77771399.6− Results for all 18 Partitions

18. 18 500 foot radius MethodMilesHours Best Time Best Distance RS3798.1165:4100 Greedy A3045.2140:0577 Greedy B3140.3144:4100 IP13055.6140:3745 IPD13039.1140:0278 Results for all 18 Partitions

19. 19  To provide redundancy, we test how serving each customer by at least two different road segments effects the costs  In terms of the IP, change to Redundancy 500 foot radius MethodMilesHours Number of Segments Miles of Segments Deadhead Miles IP2192.38:2325081.2111.1 IPD2193.18:2625179.9113.2 IP1188.28:1821678.5109.7 IPD1188.48:1821678.3110.1 Sparse Partition

20. 20  We have shown several heuristics for solving this new class of problems  The best heuristics seem to work well  RFID travel paths have a 15% time savings and 20% distance savings over the RS solution  As the technology improves (i.e., the radius increases) the savings will increase dramatically Conclusions