SolvLRDE

Release: Beta — suitable for trial and evaluation; APIs may change before GA

Laplace-resolvent solver for linear ODE/PDE/DAE systems

SolvLRDE integrates linear ODE, PDE, and DAE systems by Talbot contour Laplace inversion of the resolvent operator — direct evaluation at time T with cost independent of T and immunity to stiffness (for problems in scope).

Sold as per-vertical kernel SKUs (Grid, Climate, Control, Pricing, Thermal, Specialty). Use SolvLRDE for linear systems; use SolvSRK for nonlinear ODEs.

What it does

Many engineering problems are linear time-dependent systems — heat diffusion, power-grid transients, LTI control, pricing kernels — but teams time-step them because that is what their toolchain knows. Time-stepping cost grows with horizon T and collapses on stiffness.

SolvLRDE evaluates the solution at time T directly via Talbot contour Laplace inversion. Cost is essentially independent of T and the method is stiffness-immune for problems in scope. Use SolvLRDE for linear systems; use SolvSRK for nonlinear ODEs — mixing them is the most common integration mistake.

When to use it

Grid operators, quant pricing, climate/energy linear PDE discretizations, LTI control — anywhere you currently march a linear system forward in time and pay for it.

See Examples for runnable code.

Who it is for

• Grid operators solving power-flow and transient stability problems
• Climate and energy teams needing T-independent linear PDE evaluation
• Quant finance teams evaluating resolvent/pricing kernels (Pricing SKU)

What ships

• C library (libsolvlrde) — LAPACK-backed
• Python wheel (solvlrde)
• Dense and banded system solvers plus DAE reduction

Related products

Nonlinear dynamics: SolvSRK. Pricing UQ: SolvSRK-UQ.