Method and apparatus for providing large scale vehicle routing

1. A method for large scale vehicle routing comprising:

receiving a plurality of plans, wherein a plan of the plurality of plans assigns a vehicle, a driver of the vehicle, or a combination thereof a set of rides to traverse;

constructing a ride graph to represent feasible connections between the set of rides in a first plan of the plurality of plans, a second plan of the plurality of plans, or a combination thereof, wherein the feasible connections are connections in which the vehicle following the first plan, the second plan, or a combination can arrive from a drop-off location of a first ride to a pick-up location of a second ride without being late to a pick-up time of the second ride;

clustering the plurality of plans into one or more clusters based on a proximity measure, wherein the proximity measure indicates a proximity of the first plan to the second plan based the ride graph;

for each cluster of the one or more clusters, separately computing a solution to a large scale vehicle routing problem for the set of rides in said each cluster, wherein the large scale vehicle routing problem is a non-deterministic polynomial-time hard problem; and

causing the vehicle, the driver, or a combination thereof to traverse the set of rides based on the solution.

2. The method of claim 1 , wherein the proximity of the first plan to the second plan is based on a probability that a first driver following the first plan can take on a ride from the second plan.

3. The method of claim 1 , further comprising:

dropping one or more of the feasible connections from the ride graph, wherein the one or more feasible connections have a travel time, a distance, a wait time, or a combination thereof between rides exceeding a threshold, wherein the proximity measure is based the ride graph without the one or more feasible connections.

4. The method of claim 1 , wherein each node of the ride graph corresponds to a ride of the set of rides, the method further comprising:

constructing a first matrix representing the nodes of the ride graph of the first plan; and

constructing a second matrix representing the nodes of the ride graph of the second plan,

wherein the proximity measure is based on a distance metric calculated between the first matrix and the second matrix.

5. The method of claim 4 , wherein the distance metric is based on one or more common nodes between the first matrix and the second matrix.

6. The method of claim 5 , wherein the distance metric is based on one or more edge values between the nodes, and wherein the one or more edge values are based on whether the feasible connections are present between the nodes.

7. The method of claim 6 , further comprising:

calculating one or more variables based on the one or more common nodes,

wherein the one or more variables include at least one of:

a first variable based on the one or more edge values corresponding to first edges entering the one or more common nodes;

a second variable based on the one or more edge values corresponding to second edges exiting the one or more common nodes;

a third variable based on the one or more edge values corresponding to third edges entering the one or more common nodes from the first plan and exiting the one or more common nodes to the second plan; and

a fourth variable based on the one or more edge values corresponding to fourth edges entering the one or more common nodes from the second plan and exiting the one or more common nodes to the first plan, and

wherein the distance metric is based on the one or more variables.

8. The method of claim 1 , further comprising:

iteratively split the one or more clusters until a corresponding cluster size is below a maximum cluster size threshold, and

wherein the set of rides include at least 100 rides that comprise one or more shared rides.

9. The method of claim 1 , further comprising:

iteratively merging the one or more clusters until a corresponding cluster size is above a minimum cluster size threshold.

10. The method of claim 9 , wherein the one or more clusters are merged based a closest cluster determined using the proximity measure.

11. An apparatus for providing large scale vehicle routing, comprising:

a processor; and

a memory including computer program code for a program,

the memory and the computer program code configured to, with the processor, cause the apparatus to perform at least the following,

receive a plurality of plans, wherein a plan of the plurality of plans assigns a vehicle, a driver of the vehicle, or a combination thereof a set of rides to traverse;

construct a ride graph to represent feasible connections between the set of rides in a first plan of the plurality of plans, a second plan of the plurality of plans, or a combination thereof, wherein the feasible connections are connections in which the vehicle following the first plan, the second plan, or a combination can arrive from a drop-off location of a first ride to a pick-up location of a second ride without being late to a pick-up time of the second ride;

cluster the plurality of plans into one or more clusters based on a proximity measure, wherein the proximity measure indicates a proximity of the first plan to the second plan based the ride graph;

for each cluster of the one or more clusters, separately compute a solution to a large scale vehicle routing problem for the set of rides in said each cluster, wherein the large scale vehicle routing problem is a non-deterministic polynomial-time hard problem; and

cause the vehicle, the driver, or a combination thereof to traverse the set of rides based on the solution.

12. The apparatus of claim 11 , wherein the proximity of the first plan to the second plan is based on a probability that a first driver following the first plan can take on a ride from the second plan.

13. The apparatus of claim 11 , wherein each node of the ride graph corresponds to a ride of the set of rides, and the apparatus is further caused to:

construct a first matrix representing the nodes of the ride graph of the first plan; and

construct a second matrix representing the nodes of the ride graph of the second plan,

wherein the proximity measure is based on a distance metric calculated between the first matrix and the second matrix.

14. The apparatus of claim 13 , wherein the distance metric is based on one or more common nodes between the first matrix and the second matrix.

15. A non-transitory computer-readable storage medium for providing large scale vehicle routing, carrying one or more sequences of one or more instructions which, when executed by one or more processors, cause an apparatus to at least perform the following steps:

receiving a plurality of plans, wherein a plan of the plurality of plans assigns a vehicle, a driver of the vehicle, or a combination thereof a set of rides to traverse;

constructing a ride graph to represent feasible connections between the set of rides in a first plan of the plurality of plans, a second plan of the plurality of plans, or a combination thereof, wherein the feasible connections are connections in which the vehicle following the first plan, the second plan, or a combination can arrive from a drop-off location of a first ride to a pick-up location of a second ride without being late to a pick-up time of the second ride;

clustering the plurality of plans into one or more clusters based on a proximity measure, wherein the proximity measure indicates a proximity of the first plan to the second plan based the ride graph;

for each cluster of the one or more clusters, separately computing a solution to a large scale vehicle routing problem for the set of rides in said each cluster, wherein the large scale vehicle routing problem is a non-deterministic polynomial-time hard problem; and

causing the vehicle, the driver, or a combination thereof to traverse the set of rides based on the solution.

16. The non-transitory computer-readable storage medium of claim 15 , wherein the proximity of the first plan to the second plan is based on a probability that a first driver following the first plan can take on a ride from the second plan.

17. The non-transitory computer-readable storage medium of claim 15 , wherein each node of the ride graph corresponds to a ride of the set of rides, and the apparatus is caused to further perform:

constructing a first matrix representing the nodes of the ride graph of the first plan; and

constructing a second matrix representing the nodes of the ride graph of the second plan,

wherein the proximity measure is based on a distance metric calculated between the first matrix and the second matrix.

18. The non-transitory computer-readable storage medium of claim 17 , wherein the distance metric is based on one or more common nodes between the first matrix and the second matrix.

19. The non-transitory computer-readable storage medium of claim 18 , wherein the distance metric is based on one or more edge values between the nodes, and wherein the one or more edge values are based on whether the feasible connections are present between the nodes.

20. The apparatus of claim 14 , wherein the distance metric is based on one or more edge values between the nodes, and wherein the one or more edge values are based on whether the feasible connections are present between the nodes.