End-to-End Framework

Code as you model or model then code. Load, set, simulate, analyze, or save data where needed. Solve, generate reports, and analyze decision problems within a unified workflow using Pythonic syntax.



from feloopy import data_toolkit, search

dt = data_toolkit()
c = dt.store("c", [2, 5])
A = dt.store("A", [[5, 3], [2, 7]])
b = dt.store("b", [10, 9])

def example(m):
    x = m.ivar("x")
    y = m.pvar("y") 
    m.obj(c[0]*x+c[1]*y)
    m.con(A[0][0]*x+A[0][1]*y <= 10)
    m.con(A[1][0]*x+A[1][1]*y <= 9)
    return m

search(environment=example, directions=["max"]).clean_report()



#MILP
def example(m):
    x = m.ivar("x")
    y = m.pvar("y") 
    m.obj(c[0]*x+c[1]*y)
    m.con(A[0][0]*x+A[0][1]*y <= 10)
    m.con(A[1][0]*x+A[1][1]*y <= 9)
    return m

#MINLP
def example(m):
    x = m.ivar("x")
    y = m.pvar("y") 
    m.obj(c[0]*x**3+c[1]*y)
    m.con(A[0][0]*x**(-1)+A[0][1]*y <= 10)
    m.con(A[1][0]*x+A[1][1]*y**2 <= 9)
    return m

Versatile Toolbox

Make any certain or uncertain, convex or non-convex, single- or multi/many-objective decision problem stably solvable with LP, NLP, IP, MIP (i.e., MILP, and MINLP), or GPP (i.e., heuristics). Conclude from opinions with MADM or satisfy expectations with CP!

Configurable. Vendor-Neutral.

Easily switch between interfaces/solvers, methods/approaches, by modifying arguments. Use them together, in sequence, or in parallel—no re-coding needed! Combine the power of Julia and Python seamlessly with a low-code, prompt-free experience!


search(environment=example, 
       method="exact", 
       interface="interface_name", #e.g., "pyomo", "pulp", or "jump".
       solver="solver_name", #e.g., "cbc", "glpk", or "highs".
       approach="approach_name",  #e.g., "ecm", or "nwsm".
       options={"option_1": "value_1"}, #e.g., {"epoch": 100}.
       verbose=False, #show/hide solver log
       directions=["max"]).clean_report()


#Boilerplate benchmark

#Scenario 1
search(environment=example, 
       param_a=3, param_b=2, 
       directions=["max"]).clean_report()    

#Scenario 2
search(environment=example, 
       param_a=4, param_b=4, 
       directions=["max"]).clean_report()    

#Comprehensive benchmark
search(environment=example, 
       param_a=3, param_b=2, 
       directions=["max"], benchmark="all").clean_report()

Advanced Benchmarking

Create boilerplates of a model/algorithm and implement various tests to analyze and simulate the impact of the hyperparameters, data, and even the used interfaces and solvers! Assess model generation time, solution generation time, and the results in various scenarios to ensure a scalable impact.

Projects. Redefined.

Automate project management with intuitive CLI commands. Easily set up, build, back up, and recover FelooPy projects and manage their dependencies.


                        
flp project --name=sample-project

flp clean

flp backup

flp recover

Take FelooPy IDE with you

Join a Growing Community of users!

You may install and use FelooPy IDE like other Python libraries as follows:

Or you may use its dedicated application:

Detecting OS...