Terms

A 3D+ point is a point with latitude, longitude, elevation, and time components.

A 3D+ vector describes the angle and magnitude of the difference between two 3D+ points.

A route is a series of 3D+ points which connect to form a continuous path.

Algorithm

Score routes against given 3D+ points origin and destination

DEFINE drudge, a factor to convert lat/lng distances to a score

DEFINE wander, a factor to convert deviation from course to a score

DEFINE effort, a factor to convert elevation gain to a score

DEFINE patience, a factor to convert time to a score

DEFINE rush, a factor to convert lateness to a score

SET speed = 13mph

FOR each route

SET score to 0

FIND start, the nearest 3D+ point to origin

FIND finish, the nearest 3D+ point to destination

CREATE ride, a 3D+ vector using start and finish

CREATE commute, a 3D+ vector using origin and destination

FOR each 3D+ point step in route

GET next step

GET lean, 3D+ vector using step and next step

BUMP score by wander * difference between angle of commute and angle of lean

BUMP score by effort * elevation gain of lean

BUMP score by wander * difference between angle of commute and angle of ride * distance of ride

BUMP score by rush * (time of finish - (time of destination - distance between finish and destination / speed))

BUMP score by patience * difference between time of origin and time of start

BUMP score by patience * difference between time of finish and time of destination

SORT routes descending by score

Todo

• improve penalty for arriving late
• improve accounting for pleasant deviation
  • balance of elevation gain and loss
  • frequency of alternation between climbing and descending
  • maybe allow some configuration by the user of the conversion factors
• can I refactor comparisons to leverage 4-component vectors' inherent qualities (i.e. angle & magnitude)?