Parameter -- added waitForKey param; FastAgentsLayer -- added filled parameter
Fixed 8-agents demo
RT instance generator does not store the list of tasks anymore
fixes in RT instance generator
max speed of an agent is now represented by a float
Renamed EarliestArrivalProblem to TrajectoryCoordinationProblem which is now an umbrella interface for EarliestArrival and RelocationTask Problems.
Introduced new interface RelocationTaskCoordinationProblem. TrajectoryCoordinationProblemImpl implements both EarliestArrivalProblem and RelocationTaskCoordinationProblem.
EarliestArrivalProblem can now contain a list of taks
moved maps to a dedicated directory
Added relocation task interface and default implementation
Reverted some debugging vis.
added new corridor environment, changed visualization etc.
Solver now recognizes parameter -samplinginterval
When the penalty function returns Infinity, the edge remains in the space-time graph.
Started using ConstantStepSegmentedTrajectory to improve performance of finding position for a particular timepoint.
Minor changes to visualization
chokepoint and four-boxes environments for r=75
PenaltyMethod uses random sequence of robots
new envs and d-envs
made some of the layers togglable
samplingInterval set to 10
minonr bug fix
The samplingInterval has been set inside each algorithm. Pulled out and it is now set centrally in main().
Introduced minPenalty, added some more printouts
Minot changes to visualization
ProblemInstanceDesigner -- second argument can be used to specify the output file, agent start and goal positions snap to graph.
Moved a method from trajectory tools.
deleted outdated environments, added new ones
Solver now handles OutOfMemoryError
fixed jUnit tests
Summary contains the status at the termination of the algorithm -- SUCCESS, FAIL or TIMEOUT.
When the -timestep parameter is provided, the edges of the graph are broken into smaller edges that take exactly timestep to travel. This allows to use OD-C^n on the graph.
minor fix in the jUnit test
a script that prepares data for GNG generator
Added new environments
Fixed broken JUnit tests.
preparing the code for experiment
refactored the design of solver
introduced -waitmove parameters to the Solver
kDPM made working again, added support for -bgimg, switched to discretized time dimension
TrajectoryOptimizer returns also the penalty of the optimal trajectory
AStarTrajectoryOptimizer now supports constant discretization of the time dimension.
Added warehouse environment
added willow garage environment
Added infrastructure checker tool that can be used if a particular environment is a valid infrastructure.
VisUtil.visualizeProblem now draws the planning graph as well.
eclipsee run configurations
visualization of environment changed, boundary is not filled anymore
added new empty-hall environment and its discretized variants, fixed dock position in ubremen-r27-docks discretized environment
added agent number to mission visualization
a bug fix
fixed ubremen docks environment
ProblemDesigner can now extend existing problems. In particular, one can click on the nodes of the planning graph to specify the docks.
Adding support for creating roadmaps with infrastructure property in the triangulation generator
Instance generator now supports flag -sgnoverlap that asserts that start and goal positions of the robots do not overlap
Docks visualization fix
Problem designer can be used to add docks to the problem.
log level set to WARN
TriangulationGenerator can generate infrastructures.
Added mission visualization.
addded timestep to Parameters
Triangulation generator can now generate sparse docks
customized the visualization
radius_grace = 1.3
Merge with 870c1e254ff01d9664e9c4bf2407d190707861f0
Backed out of changeset 216:ae72eb1e2823
Added radius grace to conflict generator.
XML deserializer reports on the size of the planning graph
Merge with 76dde0ef7b90fb367431f176b8c150a69d82d3ee