Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 24, 2024

• Updated README.md with testing build/run commands

  mchaveroche

  authored
  90e7df8

Commits on Aug 4, 2023

• Update README.md

  mchaveroche

  authored
  20dd991

• Release of version 1 of efficient-DST from a previous commit,

  Show description for 12214be
  

  Maxime Chaveroche

  committed
  12214be

Commits on Apr 10, 2022

• Fixed stack problem in benchmarking: transformed the subsets array into a vector for heap allocation instead of stack. Rewrote plot.py to display results. Benchmarking done up to N=26.

  

  Maxime Chaveroche

  committed
  59062fb

Commits on Apr 7, 2022

• Removed comments from mobius_inversion. Fixed update of semilattice scheme when adding a normalizing set. Made benchmark an include class. Made demo generate a random decomposition.

  

  Maxime Chaveroche

  committed
  e43c34a

Commits on Apr 6, 2022

• Fixed lattice demo computations for decompositions,

  Show description for 4ee2470
  

  Maxime Chaveroche

  committed
  4ee2470

Commits on Mar 28, 2022

• Works, cleaner, but no performance gain...

  

  Maxime Chaveroche

  committed
  d5dceea

Commits on Mar 20, 2022

• Benchmark done

  

  Maxime Chaveroche

  committed
  7dbf64b

Commits on Mar 13, 2022

• Added vector versions of all DST functions.

  Show description for 592337f
  

  Maxime Chaveroche

  committed
  592337f

Commits on Mar 6, 2022

• Added reduced_FMT as powerset_btree-based algorithm to be able to compare

  Show description for e3ca36f
  

  Maxime Chaveroche

  committed
  e3ca36f

Commits on Feb 28, 2022

• Replaced class belief_function with a boolean check is_a_belief_function in implicability_function

  Show description for 010afa0
  

  Maxime Chaveroche

  committed
  010afa0

Commits on Feb 26, 2022

• Fixed disjunctive_decomposition and implicability_function + Removed redundant normalization in mass_function + Placed a boolean template parameter to specify if we want a classic decomposition or …

  Show description for 0efe69c
  

  Maxime Chaveroche

  committed
  0efe69c

Commits on Feb 24, 2022

• Moved normalization function to zeta_transform

  

  Maxime Chaveroche

  committed
  1921474

• Fixed decomposition normalization + Renamed DST functions

  

  Maxime Chaveroche

  committed
  68cc27a

• Fixed superset search in powerset_btree + normalization in commonality_function when taking a conjunctive_decomposition as input

  

  Maxime Chaveroche

  committed
  b902b70

Commits on Feb 22, 2022

• Conjunctive/disjunctive decompositions working. Direct computation from simple mass functions to global mass function OK. Still, there remains some errors when assigning simple masses directly and …

  Show description for 0d42435
  mchaveroche

  committed
  0d42435

Commits on Feb 11, 2022

• Fixed numerous issues.

  mchaveroche

  committed
  359fab1

Commits on Feb 9, 2022

• Fixed insert error in powerset_btree: leaf would not change when navigating through the tree. Remains a bug when computing conjunctive weights, which is at least also a bug in powerset_btree, since…

  Show description for b2f8ea4
  

  Maxime Chaveroche

  committed
  b2f8ea4

Commits on Feb 7, 2022

• fod replaced by a simpler and clearer class called sample_space.

  Show description for 1a36423
  

  Maxime Chaveroche

  committed
  1a36423

Commits on Feb 2, 2022

• Found a proper way to write a general template for all transformations while being usable. Compilations errors dealt with. It remains to switch template parameters to add the default value type flo…

  Show description for 854aa96
  mchaveroche

  committed
  854aa96

Commits on Jan 21, 2022

• Removed computation_scheme, replaced it with the template efficient_mobius_inversion in mobius_inversion, which is an adaptation of mobius_inversion_template. Removed the 8 pre-built mobius inversi…

  Show description for ce63bdf
  mchaveroche

  committed
  ce63bdf

Commits on Jan 17, 2022

• Removed the need for strict supersets search in pignistic_probability. Added insert or update if not null method in powerset_btree. Moved iota_elements logic directly into mobius_inversion_template…

  Show description for 40354ae
  mchaveroche

  committed
  40354ae

Commits on Jan 10, 2022

• Updated conjunctive and disjunctive decompositions to avoid use of strict subset or superset search + Added separate class iota_elements as they have inherent properties such as join or meet irredu…

  Show description for 930a893
  mchaveroche

  committed
  930a893

Commits on Dec 23, 2021

• insert method of powerset_btree rewritten with bitwise operations. Should be much faster. It is also more compact, much more clear and commented.

  mchaveroche

  committed
  bb3fa48

Commits on Dec 20, 2021

• Optimizing powerset_btree: now includes an unordered_map named manifest, in order to get constant time accessing + replacing bitset methods usage with bitwise operations (subsets_of and supersets_o…

  Show description for 9b6c4d4
  mchaveroche

  committed
  9b6c4d4

Commits on Dec 11, 2021

• Removed unused makefile.

  mchaveroche

  committed
  2923406

Commits on Nov 18, 2019

• Added a new version of the FMT algortihm that leverages the focal core and is fairer for benchmark both for the FMT and the EMT since it uses the same data structures.

  mchaveroche

  committed
  4e2e19d

Commits on Nov 17, 2019

• Switched to benchmark in main. Commented timers.

  mchaveroche

  committed
  4406bbc

• Set, subset and superset searches of powerset_btree rewritten with bitset search methods. Same done with powersets fusion. Works.

  mchaveroche

  committed
  f51f994

• Fixed an error in set_emptyset of conjunctive_weight. Rewrite of insert method of powerset_btree using _Find_first() and _Find_next() methods of bitset. Works. Adapted the rest of demo to the new API.

  mchaveroche

  committed
  437b0d5

Commits on Nov 12, 2019

• Attempt to optimize bitset operations. So far, changed get_final_element_number.

  mchaveroche

  committed
  89f7cec

Commits on Nov 8, 2019

• Removed dependence to to_ulong. Now, efficient_DST can truly deal with pseudo-infinite powersets.

  mchaveroche

  committed
  53ebf29

Commits on Nov 5, 2019

• Data structure changed from dynamic_bitset to bitset + Benchmark set up + plot script written

  mchaveroche

  committed
  208ed00

Commits on Oct 28, 2019

• Fixes and optimizations

  mchaveroche

  committed
  c872910

Commits on Oct 26, 2019

• sort cardinalities method moved from powerset_btree to fod. Data structure for computation schemes moved from powerset_btree to mainly unordered_map. Algorithms simplified.

  mchaveroche

  committed
  23e22a9

Pagination

PreviousNext