s14h
/
prism-accumulation
1
0
Fork 0
Code Releases Activity
3405 Commits
5 Branches
0 Tags
31 MiB
Tree: f33dea0f43
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'f33dea0f43'
${ noResults }
Commit Graph

117 Commits (f33dea0f43bb5ca39886bfbf2259f80b68ec1934)

Author SHA1 Message Date
Dave Parker f0a486b0c7 Further refactoring in explicit model checkers: push more duplicated code into ProbModelChecker.
12 years ago
Dave Parker 36997ee08c Refactor explicit model checkers a bit, including changes to way min/max info is passed around (should generalise to games more nicely).
12 years ago
Dave Parker dfce095d45 Make sure that some PrismFileLogs are close()ed after use - spotted as a bug by Bruno.
12 years ago
Dave Parker 2306ee39ac Bugfix: latest fix to policy iteration broke a previous regression test (rewpoliter.nm) - solved by doing inner solution using lowe level method which takes infinity-states as an argument (in any case, precomputation should not be needed for inner solution).
12 years ago
Dave Parker 8291b5984c Refactoring wrt the way that relational operators are stored for P/R/S operators (String -> RelOp).
12 years ago
Dave Parker d81ef64ff2 Code refactoring (checking for LTL with time bounds).
12 years ago
Dave Parker 6073712dc9 Bugfix in policy iteration (bug found by Aaron Bohy): needs to be passed in strategy from Prob1.
12 years ago
Dave Parker 4d3d52edaf Policy iteration for reachability reward problems (explicit engine).
12 years ago
Dave Parker 99a9c85a96 Fixes/tidies of explicit LTL model checking (mostly from Joachim Klein): DRA products correctly include all initial states, and probabilities are ampped back correctly.
12 years ago
Dave Parker edcac4a4a2 Added R[C<=k] operator for MDPs (sparse, explicit).
12 years ago
Dave Parker ad7e014079 Fix modified policy iteration (and add internal option to explicit state model checkers to not flag non-convergence as an error).
12 years ago
Dave Parker c68d4a4946 Refinements to PrismComponent interface, and to explicit.StateModelChecker handling of settings (originally due to bug caused by calling overridable methods in explicit model checking constructors).
13 years ago
Dave Parker 6f5213111b Fix explicit MDP model checking (EC computation was incorrect) + some refactoring.
13 years ago
Dave Parker bce5d3f3be MDP-LTL model checking for explicit (still needs fixing) + correction to DTMC-DRA product construction.
13 years ago
Dave Parker 3e2efc21e9 New PrismComponent class: refactoring for various model checking components.
13 years ago
Dave Parker ff11760860 Update strategy generation in explicit engine to distinguish between: -1 (unknown), -2 (arbitrary), -3 (unreachable).
13 years ago
Dave Parker 3b6c2ff51f Generate strategy storage for expected rewards (explicit engine).
13 years ago
Dave Parker 1c42335bcf Some refactoring of explicit model classes, to introduce NondetModel interface.
13 years ago
Dave Parker 6928edbe8c Add -exportstrat switch (currently operates in parallel to -exportadv switch).
13 years ago
Dave Parker 951f0bd507 First bits of code for improved strategy generation.
13 years ago
Dave Parker 5a1186c177 Strategy synthesis for reach rewards in the explicit engine: choices for inf states and do export.
13 years ago
Dave Parker 65a6464b32 Strategy synthesis for reach rewards in the explicit engine (no precomputation yet).
13 years ago
Dave Parker d081fdae7d Remove debug output.
13 years ago
Dave Parker 7805084dc6 Fix strategy generation for Prob1 (explicit): do an extra inner fixed point loop at the end, to avoid problems of generating surplus strategy info during early outer iterations.
13 years ago
Dave Parker 48fd9e7bdd Log output tweaks.
13 years ago
Dave Parker f404f8e8d1 Remove debug output.
13 years ago
Dave Parker 22b6e4f5ba Bug fix in previous commit.
13 years ago
Dave Parker b8969df87e MDP reachability in explicit engine: do not do numerical solution unless needed.
13 years ago
Dave Parker fee5b70ac1 Connect policy iteration algs (explicit engine) to strategy generation. Also, only export DTMC to adversary file (currently).
13 years ago
Dave Parker 3ff7df7b0a Attach strategy generation for MDP Gauss-Seidel (explicit).
13 years ago
Dave Parker 8b6d5d17a8 Improvements to strategy generation for MDP reachability in explicit engine, notably algorithms for precomputation. Also split Prob1 into Prob1A and Prob1E for efficiency reasons.
13 years ago
Dave Parker 51bf20f1f6 Add strategy generation to MDP Gauss-Seidel (explicit).
13 years ago
Dave Parker 064356d11f Rename adversary to strategy in explicit MDP model checking + align with prism-games.
13 years ago
Dave Parker 3f56c6668f Miscellaneous code changes/tidies - trying to align with prism-games a bit.
13 years ago
Dave Parker 84f1c97413 Code tidy.
13 years ago
Dave Parker 3b59b8f6cd Code tidy.
13 years ago
Dave Parker da50f86281 Tweak comments in policy iteration.
13 years ago
Dave Parker eb34f465a7 Bounded until for DTMC/MDP in explicit engine.
13 years ago
Dave Parker 433c3a3414 Next operator for explicit model checker.
13 years ago
Dave Parker 4413259325 Explicit model checker can handle negated path operators like G.
14 years ago
Dave Parker ab2d4d52c6 Bug fix: only process adversary if generated (explicit).
14 years ago
Dave Parker 17c9691d8f Adversary generation for MDPs in explicit engine restricts to reachable states.
14 years ago
Dave Parker 750f3fd0ba Separate setting explicit engine for MDP solution method (stoch games still uses old solnMethod field).
14 years ago
Dave Parker 5533d55183 Consistency in output msgs (explicit MDP engine).
14 years ago
Dave Parker a869bbeec0 Gauss-Seidel val iter for cumulative rewards in explicit engine.
14 years ago
Dave Parker e13f09f0ca Explicit model checkers should throw an exception when iterative numerical methods do not converge within maxIters.
14 years ago
Dave Parker 5dab19e7dc Add check for trying val iter from above for max probs - does not work.
14 years ago
Dave Parker 6a046b4431 Tidy/improve passing of settings to explicit model checkers.
14 years ago
Dave Parker 3351288f3a StateValues has an associated model in explicit engine; improved printing.
15 years ago
Dave Parker 14025e109d Improvements to StateValues class + it's use in explicit model checking.
15 years ago