Identify a convex solver that consistently handles the mixed piecewise-smooth AMM problems

Develop or identify a convex optimization solver or algorithmic variant that can be relied upon to consistently perform well on the AMM arbitrage convex programs that combine smooth and piecewise-linear constraints, including markets dominated by narrow levered curves and limit-order-like segments.

Background

Because levered and limit-order-like AMMs induce piecewise-smooth constraints, the authors report inconsistent performance from standard convex solvers. They emphasize the practical need for a solver that performs robustly across corner and interior solution regimes.

A dependable solver would enable the convex formulation to be used as a production method for arbitrage and routing in markets with concentrated liquidity.