To string added to SIPPNode
Changed the way polygons are visualized. Now only the border is shown.
Added support for polygons that are filled outside.
Added default bounds
Problem with generating full for SIPP temporaly fixed
print help made public in Solver
SIPPDebug class introduced
uncomitted files from previous commit
SIPP renamed to PP_SIPP
alpha version of SIPP
Local Solver parsed from Solver
minor refactoring in AlgorithmODCN
minor fix in AbstractAlgorithm
additional points extension point now generates full graph
Early functional implementation of SIPP
minor TYPO bug fixes
compilation error fixed
not tested yet
CSAIL map works
superconflict for illustration
Fixed the DPM to support k=2
Removed visio from instance generator
Fixes and improvements of the DPM planner
KSFSO renamed to KDPMD
Gradient Optimizer and kDPM-C kind of work
introduced time to penalty function
Fixed broken parts of the code. IIHP is currently broken.
GradientOptimizer seems to work
Added progress visualization. Introduced tan multiplier.
Problem: crashed if the start and target points ware the sameSolution: introduced new type of trajectory for stationary agents
Added a new simple instance generator that can generate dense regular grid instances
Merge with 715b9aad3451fcc3194257ea8cd337472774170b
Optimized the behaviour of the conflict generator
Solver method changed to work with UnifiedSolver
Solve(algorithm) function extracted
Added hard constraint phase to SFO planner, SFO uses BumpFunction and refactorizations
Added support for L1 and L2 heuristics
KSFO and PP now generate trajectories up to maxtime.
Fixed summary outputs
KSFO -k 1 was sometimes willing to accept high penalties instead of returning solution non-existence. Fixed.
Added -k parameter to the solver that controls the number of iterations performed by KSFO
Drastic refactoring of the separable optimal flow optimizer (iterative planner) work in progress
Conflict generator can be given a rectangular region in which the conflict should be located.
Conflict generator improved
- now it generates a start and target points ONLY ON the predefined grid- all agents share the very same grid- code cleaned a bit
few minor fixes
1) Vizualization now shows the algorithm name2) OD now has a option to stop3) All the edges returned by od has a fixed time duration
problem with visualization of PP fixed
PP was throwing UnsuportedOperation exceptions while trying to visualize motion graph
Fixed problem with paths
All paths shorter than a maximum length (in the number of edges) had unnecessary zero length edges as it suffix.
Fixed problem with subobtimality.
Algorithm did not stop exploring the search space subspace of agent after finding its state to be a goal state.
Time limit was not properly implemented in OD solvere. Also null pointer exception was thrown if the algorithm did not found solution.
OD renamed to ODCN
testsuite and instance generator uses the name ODCN
OD_PIN renamed to ODPIN
testsuite and instance generator uses the name ODPIN
Prioritized Planning uses perfect heuristics
added setter to prioritized planning solver just to add the option to use different heuristics
Fixed visualization exceptions
Merge with b103e25ca12afee6a5ae8eab5905f5b4a2c8c740
Merge with e2234922273a1875c97911cbe45158d5a35e6eb7
New solver replaced old one
and also Enum containing implemented algorithms was extracted. Test for OD introduced.
super class solver was refactored into multiple of simple classes
ODCN solver implemented