About

sciFLT is a Fuzzy Logic Toolbox for scilab. The toolbox was wrote in a mix of C, Fortran, TK/TCL and scilab Code.

sciFLT was fully tested under Windows and Linux, both using Scilab 3.0

sciFLT have the following properties:

• Types of Fuzzy Logic systems: At this stage, sciFLT can deal with Takagi-Sugeno Fuzzy and Mamdani fuzzy systems usage.
  
• SCICOS support: Member functions, S-Norm, T-Norm, Complement and Fuzzy Logic System (fls) are supported. Also include a palette.
• S-Norm Class supported: Dubois-Prade, Yager, Drastic sum, Einstein,  sum, Algebraic sum, Maximum.
• T-Norm Class supported: Dubois-Prade, Yager, Drastic product, Einstein product,  Algebraic product,  Minimum
• Complement Class supported:  One (classic), Yager, Dubois.
• Implication method supported: Product, Minimum
  
• Aggregation method supported: Algebraic Sum, Maximum
  
• Defuzzification method supported: Centroide, Bisector, Mean of Maximum, Largest of Maximum, Smallest of Maximu, Weighted Average, Center Average.
• Member functions supported:  Triangular, Trapezoidal, Gaussian, Extended Gaussian,  Sigmoidal,  Product of two Sigmoidal,  Difference of two sigmoidal, S-Shaped, Z-Shaped, Pi-Shaped. Linear and Const as output for Takagi-Sugeno type.
  
• GUI editor: Built in editor to change properties, inputs/outputs variables, member functions, rules.
  

Screen shots

Future plans

• Fuzzy Clustering.
• Learning Algorithm.

Todo

• Improve fls editor (buggy).
• Better integration with SCICOS, also add more blocks (as an example tune consequents).
  
• Get better performance in computational routines (as an example unroll loops).
• Compile with LCC in Windows? (see builder.sce)
  

Download