Revert "Add a bunch of tests at different scales. Switch to using linear acceleration estimates to add penalties to start/end of trips."

This reverts commit 6b9006f6e7.
This commit is contained in:
Daniel Patterson 2019-02-10 00:42:08 -08:00
parent 44d4903985
commit 6b961dccfc
No known key found for this signature in database
GPG Key ID: 19C12BE1725A028B
3 changed files with 73 additions and 215 deletions

View File

@ -0,0 +1,36 @@
@routing @maxspeed @car
Feature: Car - Acceleration profiles
Background: Use specific speeds
Given the profile "car"
Given a grid size of 1000 meters
Scenario: Car - Use stoppage penalty at waypoints
Given the node map
"""
a 1 2 3 4 5 b
"""
And the ways
| nodes | highway | maxspeed:forward | maxspeed:backward |
| ab | trunk | 60 | 45 |
And the query options
| stoppage_penalty | 5,100 |
When I route I should get
| from | to | route | time |
| a | b | ab,ab | 481.1s |
When I route I should get
| from | to | route | time |
| b | a | ab,ab | 609.9s |
When I route I should get
| from | to | route | time |
| a | a | ab,ab | 0s |
When I request a travel time matrix I should get
| | a | b |
| a | 0 | 481.1 |
| b | 609.9 | 0 |

View File

@ -3,102 +3,9 @@ Feature: Testbot - Acceleration profiles
Background: Use specific speeds
Given the profile "testbot"
Scenario: Testbot - No stoppage penalties
Given a grid size of 10 meters
Given the node map
"""
a 1 2 3 4 5 b
"""
And the ways
| nodes | highway | maxspeed:forward | maxspeed:backward |
| ab | trunk | 60 | 45 |
When I route I should get
| from | to | route | time | distance |
| a | b | ab,ab | 3.6s | 59.9m |
| a | 1 | ab,ab | 0.6s | 10m |
| a | 2 | ab,ab | 1.2s | 20m |
| a | 3 | ab,ab | 1.8s | 30m |
| a | 4 | ab,ab | 2.4s | 40m |
| a | 5 | ab,ab | 3s | 50m |
| 5 | b | ab,ab | 0.6s | 9.9m |
| 4 | b | ab,ab | 1.2s | 19.9m |
| 3 | b | ab,ab | 1.8s | 29.9m |
| 2 | b | ab,ab | 2.4s | 39.9m |
| 1 | b | ab,ab | 3s | 49.9m |
| 1 | 2 | ab,ab | 0.6s | 10m |
| 1 | 3 | ab,ab | 1.2s | 20m |
| 1 | 4 | ab,ab | 1.8s | 30m |
| 1 | 5 | ab,ab | 2.4s | 40m |
When I request a travel time matrix I should get
| | a | 1 | 2 | 3 | 4 | 5 | b |
| a | 0 | 0.6 | 1.2 | 1.8 | 2.4 | 3 | 3.6 |
| 1 | 0.8 | 0 | 0.6 | 1.2 | 1.8 | 2.4 | 3 |
| 2 | 1.6 | 0.8 | 0 | 0.6 | 1.2 | 1.8 | 2.4 |
| 3 | 2.4 | 1.6 | 0.8 | 0 | 0.6 | 1.2 | 1.8 |
| 4 | 3.2 | 2.4 | 1.6 | 0.8 | 0 | 0.6 | 1.2 |
| 5 | 4 | 3.2 | 2.4 | 1.6 | 0.8 | 0 | 0.6 |
| b | 4.8 | 4 | 3.2 | 2.4 | 1.6 | 0.8 | 0 |
Scenario: Testbot - No stoppage points, tiny grid size
Given a grid size of 1 meters
Given the node map
"""
a 1 2 3 4 5 6 7 8 9 b
"""
And the ways
| nodes | highway | maxspeed:forward | maxspeed:backward |
| ab | trunk | 60 | 45 |
When I request a travel time matrix I should get
| | a | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | b |
| a | 0 | 0 | 0.1 | 0.1 | 0.2 | 0.3 | 0.3 | 0.4 | 0.4 | 0.5 | 0.6 |
| 1 | 0 | 0 | 0.1 | 0.1 | 0.2 | 0.3 | 0.3 | 0.4 | 0.4 | 0.5 | 0.6 |
| 2 | 0.1 | 0.1 | 0 | 0 | 0.1 | 0.2 | 0.2 | 0.3 | 0.3 | 0.4 | 0.5 |
| 3 | 0.2 | 0.2 | 0 | 0 | 0.1 | 0.2 | 0.2 | 0.3 | 0.3 | 0.4 | 0.5 |
| 4 | 0.3 | 0.3 | 0.2 | 0.1 | 0 | 0.1 | 0.1 | 0.2 | 0.2 | 0.3 | 0.4 |
| 5 | 0.4 | 0.4 | 0.3 | 0.2 | 0.1 | 0 | 0 | 0.1 | 0.1 | 0.2 | 0.3 |
| 6 | 0.4 | 0.4 | 0.3 | 0.2 | 0.1 | 0 | 0 | 0.1 | 0.1 | 0.2 | 0.3 |
| 7 | 0.5 | 0.5 | 0.4 | 0.3 | 0.2 | 0.1 | 0.1 | 0 | 0 | 0.1 | 0.2 |
| 8 | 0.6 | 0.6 | 0.5 | 0.4 | 0.3 | 0.2 | 0.2 | 0 | 0 | 0.1 | 0.2 |
| 9 | 0.7 | 0.7 | 0.6 | 0.5 | 0.4 | 0.3 | 0.3 | 0.2 | 0.1 | 0 | 0.1 |
| b | 0.8 | 0.8 | 0.7 | 0.6 | 0.5 | 0.4 | 0.4 | 0.3 | 0.2 | 0.1 | 0 |
Scenario: Testbot - No stoppage points, tiny grid size
Given a grid size of 1 meters
Given the node map
"""
a 1 2 3 4 5 6 7 8 9 b
"""
And the ways
| nodes | highway | maxspeed:forward | maxspeed:backward |
| ab | trunk | 60 | 45 |
And the query options
| stoppage_penalty | 1.85,15 |
When I request a travel time matrix I should get
| | a | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | b |
| a | 0 | 1.4 | 2 | 2.5 | 2.9 | 3.2 | 3.6 | 3.8 | 4.1 | 4.4 | 4.6 |
| 1 | 1.4 | 0 | 1.4 | 2 | 2.5 | 2.9 | 3.2 | 3.6 | 3.8 | 4.1 | 4.4 |
| 2 | 2 | 1.4 | 0 | 1.4 | 2 | 2.5 | 2.9 | 3.2 | 3.6 | 3.8 | 4.1 |
| 3 | 2.5 | 2 | 1.4 | 0 | 1.4 | 2 | 2.5 | 2.9 | 3.2 | 3.6 | 3.8 |
| 4 | 2.9 | 2.5 | 2 | 1.4 | 0 | 1.4 | 2 | 2.5 | 2.9 | 3.2 | 3.6 |
| 5 | 3.2 | 2.9 | 2.5 | 2 | 1.4 | 0 | 1.4 | 2 | 2.5 | 2.9 | 3.2 |
| 6 | 3.6 | 3.2 | 2.9 | 2.5 | 2 | 1.4 | 0 | 1.4 | 2 | 2.5 | 2.9 |
| 7 | 3.8 | 3.6 | 3.2 | 2.9 | 2.5 | 2 | 1.4 | 0 | 1.4 | 2 | 2.5 |
| 8 | 4.1 | 3.8 | 3.6 | 3.2 | 2.9 | 2.5 | 2 | 1.4 | 0 | 1.4 | 2 |
| 9 | 4.4 | 4.1 | 3.8 | 3.6 | 3.2 | 2.9 | 2.5 | 2 | 1.4 | 0 | 1.4 |
| b | 4.6 | 4.4 | 4.1 | 3.8 | 3.6 | 3.2 | 2.9 | 2.5 | 2 | 1.4 | 0 |
Given a grid size of 1000 meters
Scenario: Testbot - Use stoppage penalty at waypoints
Given a grid size of 10 meters
Given the node map
"""
a 1 2 3 4 5 b
@ -109,60 +16,21 @@ Feature: Testbot - Acceleration profiles
| ab | trunk | 60 | 45 |
And the query options
| stoppage_penalty | 1.85,15 |
| stoppage_penalty | 5,100 |
When I route I should get
| from | to | route | time |
| a | b | ab,ab | 412.3s |
When I route I should get
| from | to | route | time |
| b | a | ab,ab | 505.9s |
When I route I should get
| from | to | route | time |
| a | a | ab,ab | 0s |
When I request a travel time matrix I should get
| | a | 1 | 2 | 3 | 4 | 5 | b |
| a | 0 | 4.6 | 6.5 | 8 | 9.3 | 10.4 | 11.3 |
| 1 | 4.6 | 0 | 4.6 | 6.5 | 8 | 9.3 | 10.3 |
| 2 | 6.5 | 4.6 | 0 | 4.6 | 6.5 | 8 | 9.2 |
| 3 | 8 | 6.5 | 4.6 | 0 | 4.6 | 6.5 | 8 |
| 4 | 9.3 | 8 | 6.5 | 4.6 | 0 | 4.6 | 6.5 |
| 5 | 10.4 | 9.3 | 8 | 6.5 | 4.6 | 0 | 4.6 |
| b | 11.3 | 10.3 | 9.2 | 8 | 6.5 | 4.6 | 0 |
Scenario: Long distance grid with no penalty
Given a grid size of 1000 meters
Given the node map
"""
a 1 2 3 4 5 b
"""
And the ways
| nodes | highway | maxspeed:forward | maxspeed:backward |
| ab | trunk | 60 | 45 |
When I request a travel time matrix I should get
| | a | 1 | 2 | 3 | 4 | 5 | b |
| a | 0 | 59.9 | 119.9 | 179.9 | 239.9 | 299.9 | 359.9 |
| 1 | 79.9 | 0 | 60 | 120 | 180 | 240 | 300 |
| 2 | 159.9 | 80 | 0 | 60 | 120 | 180 | 240 |
| 3 | 239.9 | 160 | 80 | 0 | 60 | 120 | 180 |
| 4 | 319.9 | 240 | 160 | 80 | 0 | 60 | 120 |
| 5 | 399.9 | 320 | 240 | 160 | 80 | 0 | 60 |
| b | 479.9 | 400 | 320 | 240 | 160 | 80 | 0 |
Scenario: Long distance grid
Given a grid size of 1000 meters
Given the node map
"""
a 1 2 3 4 5 b
"""
And the ways
| nodes | highway | maxspeed:forward | maxspeed:backward |
| ab | trunk | 60 | 45 |
And the query options
| stoppage_penalty | 1.85,15 |
When I request a travel time matrix I should get
| | a | 1 | 2 | 3 | 4 | 5 | b |
| a | 0 | 68.9 | 128.9 | 188.9 | 248.9 | 308.9 | 368.9 |
| 1 | 86.6 | 0 | 69 | 129 | 189 | 249 | 309 |
| 2 | 166.6 | 86.7 | 0 | 69 | 129 | 189 | 249 |
| 3 | 246.6 | 166.7 | 86.7 | 0 | 69 | 129 | 189 |
| 4 | 326.6 | 246.7 | 166.7 | 86.7 | 0 | 69 | 129 |
| 5 | 406.6 | 326.7 | 246.7 | 166.7 | 86.7 | 0 | 69 |
| b | 486.6 | 406.7 | 326.7 | 246.7 | 166.7 | 86.7 | 0 |
| | a | b |
| a | 0 | 412.3 |
| b | 505.9 | 0 |

View File

@ -86,10 +86,8 @@ Status TablePlugin::HandleRequest(const RoutingAlgorithmsInterface &algorithms,
bool request_distance = params.annotations & api::TableParameters::AnnotationsType::Distance;
bool request_duration = params.annotations & api::TableParameters::AnnotationsType::Duration;
constexpr bool always_calculate_distance = true;
auto result_tables_pair = algorithms.ManyToManySearch(
snapped_phantoms, params.sources, params.destinations, always_calculate_distance);
snapped_phantoms, params.sources, params.destinations, request_distance);
if ((request_duration && result_tables_pair.first.empty()) ||
(request_distance && result_tables_pair.second.empty()))
@ -99,55 +97,20 @@ Status TablePlugin::HandleRequest(const RoutingAlgorithmsInterface &algorithms,
std::vector<api::TableAPI::TableCellRef> estimated_pairs;
// Adds some time to adjust for getting up to speed and slowing down to a stop
// Returns a new `duration`
auto adjust_for_startstop = [](const double &comfortable_acceleration,
const EdgeDuration &duration,
const EdgeDistance &distance) -> EdgeDuration {
// Assume linear acceleration from 0 to velocity
// auto comfortable_acceleration = 1.85; // m/s^2
// Very short paths can end up with 0 duration. That'll lead to a divide
// by zero, so instead, we'll assume the travel speed is 10m/s (36km/h).
// Typically, the distance is also short, so we're quibbling at tiny numbers
// here, but tiny numbers is what this adjustment lambda is all about,
// so we do try to be reasonable.
const auto average_speed =
duration == 0 ? 10 : distance /
(duration / 10.); // duration is in deciseconds, we need m/sec
// Using the equations of motion as a simple approximation, assuming constant acceleration
// https://en.wikipedia.org/wiki/Equations_of_motion#Constant_translational_acceleration_in_a_straight_line
const auto distance_to_full_speed =
(average_speed * average_speed) / (2 * comfortable_acceleration);
/*
std::cout << "Comfortable accel is " << comfortable_acceleration << std::endl;
std::cout << "Average speed is " << average_speed << " duration is " << duration
<< std::endl;
std::cout << "Distance is " << distance << " distance to full speed is "
<< distance_to_full_speed << std::endl;
*/
if (distance_to_full_speed > distance / 2)
{
// std::cout << "Distance was too short, so only using half" << std::endl;
const auto time_to_halfway = std::sqrt(distance / comfortable_acceleration);
return (2 * time_to_halfway) * 10; // result is in deciseconds
}
else
{
// std::cout << "Distance was long, using cruising speed" << std::endl;
const auto cruising_distance = distance - 2 * distance_to_full_speed;
const auto cruising_time = cruising_distance / average_speed;
const auto acceleration_time = average_speed / comfortable_acceleration;
// std::cout << "Cruising distance is " << cruising_distance << std::endl;
// std::cout << "Cruising time is " << cruising_time << std::endl;
// std::cout << "Acceleration time is " << acceleration_time << std::endl;
return (cruising_time + 2 * acceleration_time) * 10; // result is in deciseconds
}
auto penalize = [&params](float speed, EdgeDuration duration) -> EdgeDuration {
// this is the penalty before adjusting for the length of the route
auto speed_ratio = speed * MAXIMAL_ACCEL_DECEL_PENALIZABLE_SPEED_INVERSE;
auto penalty = speed_ratio * (params.max_stoppage_penalty - params.min_stoppage_penalty) +
params.min_stoppage_penalty;
// how much distance would we cover if we took half of it at this speed
// scale the penalty based on this distance, if its small we dont get much
// of the penalty, buf if its large we get up to the whole penalty
// .05 is to scale the duration to second and cut it in half
// .005 is like / 200m, so the estimated distance as a ratio of 200m
// so you get the full penalty if your route is 200m or more
auto distance_ratio = (speed * duration * .05) * 0.008;
return static_cast<EdgeDuration>(
(distance_ratio < 1.f ? distance_ratio * penalty : penalty) * 10.);
};
// Scan table for null results - if any exist, replace with distance estimates
@ -164,17 +127,16 @@ Status TablePlugin::HandleRequest(const RoutingAlgorithmsInterface &algorithms,
// apply an accel/deceleration penalty to the duration
if (result_tables_pair.first[table_index] != MAXIMAL_EDGE_DURATION && row != column)
{
/*
const auto &source =
snapped_phantoms[params.sources.empty() ? row : params.sources[row]];
const auto &destination =
snapped_phantoms[params.destinations.empty() ? column
: params.destinations[column]];
*/
result_tables_pair.first[table_index] =
adjust_for_startstop(params.min_stoppage_penalty,
result_tables_pair.first[table_index],
result_tables_pair.second[table_index]);
result_tables_pair.first[table_index] +=
penalize(source.speed_approximation,
result_tables_pair.first[table_index]) +
penalize(destination.speed_approximation,
result_tables_pair.first[table_index]);
}
// Estimate null results based on fallback_speed (if valid) and distance
if (params.fallback_speed != INVALID_FALLBACK_SPEED && params.fallback_speed > 0 &&
@ -225,14 +187,6 @@ Status TablePlugin::HandleRequest(const RoutingAlgorithmsInterface &algorithms,
}
}
// If distances weren't requested, delete them from the result so they don't
// get rendered.
if (!request_distance)
{
std::vector<EdgeDistance> empty;
result_tables_pair.second.swap(empty);
}
api::TableAPI table_api{facade, params};
table_api.MakeResponse(result_tables_pair, snapped_phantoms, estimated_pairs, result);