GitLab

Julian Zobel authored Oct 02, 2018 and Julian Zobel committed Oct 02, 2018

Fixed a bug in the weighted transition strategy where nodes would most probably just wait the minimum pause time (plus a little bit) and therefore the waiting times were not quite right (especially for big differences)


(cherry picked from commit 67d8cbd0)

- Visualization
- Refactored Route-Sensor

+ MILPv2 now uses the actual route length instead of the euclidean distance
+ Setup the environments for the next evaluation step
+ Implemented the SpotUtilizationMetric

and corresponding necessities.

~ HotspotDistribution would lead to massive (more as intended) clustering over simulation time.
~ SearchDistanceThreshold did not increase.

+ Made the node debug a standalone component similar to the Monitor class positioned properly inside the api and peerfact.
+ Added filtering capabilities
+ Added logging in some classes deep in the architecture the provide node information

+ Added the possibility to switch the transition mid-simulation for the ModularMultiTypeMovement
+ Added a new transition strategy

+ ModularMovementIconVis can now visualise nodes dependent on their states too
+ Implemented visualisation for "walking" state.

Movement:
+ Added a multi-type movement model which allows changes of the movement-type mid-simulation per host

Spot Positioning:
+ (Car)Positions of the actors can now be recorded and written to a CSV File which can then be used for the placement of the spots

Spot Assignment:
+Actors don't get a DummyLocation as SpotAssignment anymore, now actual Spots are assigned
+First assignment strategy: Shortest distance assignment

Visualisation:
+ Visualisation Layer
+ OnClick now displays the destination, the assigned Spot and the searchThreshold.