Hello SymPy Community,

I am Sai Kumar, and I propose enhancing SymPy's code generation 
capabilities to efficiently compute Jacobians and Hessians for optimization 
and ODE integration. This project will bridge the gap between symbolic and 
numerical computing, enabling users to seamlessly integrate SymPy with 
SciPy's optimization and ODE solvers.
To extend SymPy’s capabilities, I propose implementing the following 
improvements:

1 . Efficient Jacobian and Hessian Evaluation:

   - Develop functions like generate_minimize_derivative_funcs to generate 
   numerical   functions, Jacobians, and Hessians from symbolic expressions.


   - Optimize performance by reusing common subexpressions and leveraging 
   SymPy's symbolic differentiation.

2 . ODE Integration Support:

   -  Implement generate_ode_derivative_funcs to generate numerical 
   functions and       Jacobians for ODE systems.
   -  Add sparsity pattern detection for large-scale ODE systems to 
   improve                     computational efficiency.

3 . Seamless Integration with SciPy:

   - Ensure compatibility with SciPy's optimization (minimize) and ODE 
   (solve_ivp) solvers.


   - Provide clear documentation and examples for users to easily integrate 
   SymPy with SciPy.

4 . Performance Optimization:

   - Benchmark the new functionality against existing tools like pyodesys 
   and symopt.


   - Optimize code generation to minimize computational overhead for 
   large-scale problems.

These enhancements will make SymPy more versatile and efficient for 
optimization and ODE integration, benefiting researchers, engineers, and 
students working on complex numerical problems. I would love to discuss 
this further and receive feedback from the community on how best to 
approach these improvements.

Looking forward to your insights!

Best regards,
Sai Kumar




New chat

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/sympy/fd7b6f58-8a34-4ba1-9fae-7bc18cb37734n%40googlegroups.com.

Reply via email to