OK done @benmsanderson.bsky.social @jquaas.bsky.social @rustneversleeps.bsky.social
github.com/chrisroadmap...
I'll work on improving the climate model and updating the socioeconomic assumptions soon, as was done in this paper (iopscience.iop.org/article/10.1...).
github.com/chrisroadmap...
I'll work on improving the climate model and updating the socioeconomic assumptions soon, as was done in this paper (iopscience.iop.org/article/10.1...).
November 17, 2024 at 8:56 PM
OK done @benmsanderson.bsky.social @jquaas.bsky.social @rustneversleeps.bsky.social
github.com/chrisroadmap...
I'll work on improving the climate model and updating the socioeconomic assumptions soon, as was done in this paper (iopscience.iop.org/article/10.1...).
github.com/chrisroadmap...
I'll work on improving the climate model and updating the socioeconomic assumptions soon, as was done in this paper (iopscience.iop.org/article/10.1...).
i really need to figure out a way to get claude/codex to write the hamiltonian interior point filter plugin for ipopt
so i can solve the fucking singular arc problem dead...
so i can solve the fucking singular arc problem dead...
July 24, 2026 at 5:51 AM
i really need to figure out a way to get claude/codex to write the hamiltonian interior point filter plugin for ipopt
so i can solve the fucking singular arc problem dead...
so i can solve the fucking singular arc problem dead...
𝐔𝐧𝐨𝐒𝐨𝐥𝐯𝐞𝐫.𝐣𝐥 now a registered Julia package 🥳
Uno is a highly customizable #SQP and #barrier solver for nonlinearly constrained optimization. It can mimic #filterSQP (SQP method) and #IPOPT (barrier method)
👉 github.com/cvanaret/Uno
#optimization #numericaloptimization #julialang #oss #foss #cpp
Uno is a highly customizable #SQP and #barrier solver for nonlinearly constrained optimization. It can mimic #filterSQP (SQP method) and #IPOPT (barrier method)
👉 github.com/cvanaret/Uno
#optimization #numericaloptimization #julialang #oss #foss #cpp
October 30, 2025 at 2:37 PM
𝐔𝐧𝐨𝐒𝐨𝐥𝐯𝐞𝐫.𝐣𝐥 now a registered Julia package 🥳
Uno is a highly customizable #SQP and #barrier solver for nonlinearly constrained optimization. It can mimic #filterSQP (SQP method) and #IPOPT (barrier method)
👉 github.com/cvanaret/Uno
#optimization #numericaloptimization #julialang #oss #foss #cpp
Uno is a highly customizable #SQP and #barrier solver for nonlinearly constrained optimization. It can mimic #filterSQP (SQP method) and #IPOPT (barrier method)
👉 github.com/cvanaret/Uno
#optimization #numericaloptimization #julialang #oss #foss #cpp
Uno's success story #1 (by @franckgaga.bsky.social) on the inverted pendulum case study 🎉
Uno unifies SQP and barrier methods in a highly customizable framework. Try it out!
github.com/cvanaret/Uno
#optimization #mathematicaloptimization #mpc #modelpredictivecontrol #julialang #oss #foss #ipopt
Uno unifies SQP and barrier methods in a highly customizable framework. Try it out!
github.com/cvanaret/Uno
#optimization #mathematicaloptimization #mpc #modelpredictivecontrol #julialang #oss #foss #ipopt
November 18, 2025 at 12:08 PM
Uno's success story #1 (by @franckgaga.bsky.social) on the inverted pendulum case study 🎉
Uno unifies SQP and barrier methods in a highly customizable framework. Try it out!
github.com/cvanaret/Uno
#optimization #mathematicaloptimization #mpc #modelpredictivecontrol #julialang #oss #foss #ipopt
Uno unifies SQP and barrier methods in a highly customizable framework. Try it out!
github.com/cvanaret/Uno
#optimization #mathematicaloptimization #mpc #modelpredictivecontrol #julialang #oss #foss #ipopt
𝐔𝐧𝐨 has hit the 400-star mark on Github ✨
Really happy to see so much enthusiasm about our work!
Install 𝐔𝐧𝐨𝐒𝐨𝐥𝐯𝐞𝐫.𝐣𝐥 in no time, and it's plug-and-play: github.com/cvanaret/Uno...
#optimization #mathematicalprogramming #cpp #julialang #foss #oss #ipopt #modelpredictivecontrol #optimalcontrol
Really happy to see so much enthusiasm about our work!
Install 𝐔𝐧𝐨𝐒𝐨𝐥𝐯𝐞𝐫.𝐣𝐥 in no time, and it's plug-and-play: github.com/cvanaret/Uno...
#optimization #mathematicalprogramming #cpp #julialang #foss #oss #ipopt #modelpredictivecontrol #optimalcontrol
November 21, 2025 at 2:08 PM
𝐔𝐧𝐨 has hit the 400-star mark on Github ✨
Really happy to see so much enthusiasm about our work!
Install 𝐔𝐧𝐨𝐒𝐨𝐥𝐯𝐞𝐫.𝐣𝐥 in no time, and it's plug-and-play: github.com/cvanaret/Uno...
#optimization #mathematicalprogramming #cpp #julialang #foss #oss #ipopt #modelpredictivecontrol #optimalcontrol
Really happy to see so much enthusiasm about our work!
Install 𝐔𝐧𝐨𝐒𝐨𝐥𝐯𝐞𝐫.𝐣𝐥 in no time, and it's plug-and-play: github.com/cvanaret/Uno...
#optimization #mathematicalprogramming #cpp #julialang #foss #oss #ipopt #modelpredictivecontrol #optimalcontrol
Over the years, I’ve developed a unique skill: commenting on iterations of IPOPT like it’s my favorite sports team ⚽️🥅🏀
« And it went out the recovery mode !»
December 20, 2024 at 8:45 PM
Over the years, I’ve developed a unique skill: commenting on iterations of IPOPT like it’s my favorite sports team ⚽️🥅🏀
« And it went out the recovery mode !»
Just wondering how long after I go to bed it will be when this stubborn simulation finally converges… I'm guessing a few minutes…
January 31, 2024 at 2:50 AM
Just wondering how long after I go to bed it will be when this stubborn simulation finally converges… I'm guessing a few minutes…
Uno's success story #2 (by Fredrik Bagge Carlson)🔥
Uno's filtersqp preset is particularly robust on mid-size problems. Try it out! github.com/cvanaret/Uno
#optimization #mathematicaloptimization #mathematicalprogramming #sqp #barrier #oss #foss #ipopt #mpc #modelpredictivecontrol #cpp #julialang
Uno's filtersqp preset is particularly robust on mid-size problems. Try it out! github.com/cvanaret/Uno
#optimization #mathematicaloptimization #mathematicalprogramming #sqp #barrier #oss #foss #ipopt #mpc #modelpredictivecontrol #cpp #julialang
November 20, 2025 at 1:25 PM
Uno's success story #2 (by Fredrik Bagge Carlson)🔥
Uno's filtersqp preset is particularly robust on mid-size problems. Try it out! github.com/cvanaret/Uno
#optimization #mathematicaloptimization #mathematicalprogramming #sqp #barrier #oss #foss #ipopt #mpc #modelpredictivecontrol #cpp #julialang
Uno's filtersqp preset is particularly robust on mid-size problems. Try it out! github.com/cvanaret/Uno
#optimization #mathematicaloptimization #mathematicalprogramming #sqp #barrier #oss #foss #ipopt #mpc #modelpredictivecontrol #cpp #julialang
My nonconvex #optimization solver Uno (#SQP + #barrier) has reached floor(100𝝿) ⭐ on Github
Assembles building blocks (such as computation of descent directions and globalization techniques) effortlessly
github.com/cvanaret/Uno
#foss #operationsresearch #numericaloptimization #cpp #IPOPT #filterSQP
Assembles building blocks (such as computation of descent directions and globalization techniques) effortlessly
github.com/cvanaret/Uno
#foss #operationsresearch #numericaloptimization #cpp #IPOPT #filterSQP
March 10, 2025 at 8:35 PM
My nonconvex #optimization solver Uno (#SQP + #barrier) has reached floor(100𝝿) ⭐ on Github
Assembles building blocks (such as computation of descent directions and globalization techniques) effortlessly
github.com/cvanaret/Uno
#foss #operationsresearch #numericaloptimization #cpp #IPOPT #filterSQP
Assembles building blocks (such as computation of descent directions and globalization techniques) effortlessly
github.com/cvanaret/Uno
#foss #operationsresearch #numericaloptimization #cpp #IPOPT #filterSQP
Claude 5.0 is on the motherfucker - 3 million tokens in...
Part of the idea Claude was like "you dumbass that's what Ipopt does" and the other part it was like "that an exactly correct - and exciting possibility"
Part of the idea Claude was like "you dumbass that's what Ipopt does" and the other part it was like "that an exactly correct - and exciting possibility"
i really need to figure out a way to get claude/codex to write the hamiltonian interior point filter plugin for ipopt
so i can solve the fucking singular arc problem dead...
so i can solve the fucking singular arc problem dead...
July 25, 2026 at 1:07 AM
Claude 5.0 is on the motherfucker - 3 million tokens in...
Part of the idea Claude was like "you dumbass that's what Ipopt does" and the other part it was like "that an exactly correct - and exciting possibility"
Part of the idea Claude was like "you dumbass that's what Ipopt does" and the other part it was like "that an exactly correct - and exciting possibility"
𝐔𝐧𝐨 is catching up to the big names on the 𝐌𝐢𝐭𝐭𝐞𝐥𝐦𝐚𝐧𝐧 𝐛𝐞𝐧𝐜𝐡𝐦𝐚𝐫𝐤! ⚡
Uno v2.9.0 stepped up its barrier method and now solves 45/47 𝐢𝐧𝐬𝐭𝐚𝐧𝐜𝐞𝐬
github.com/cvanaret/Uno/releases/tag/v2.9.0
#optimization #mathematicalprogramming #numericaloptimization #barrier #oss #foss #julialang #unosolver #python #cpp
Uno v2.9.0 stepped up its barrier method and now solves 45/47 𝐢𝐧𝐬𝐭𝐚𝐧𝐜𝐞𝐬
github.com/cvanaret/Uno/releases/tag/v2.9.0
#optimization #mathematicalprogramming #numericaloptimization #barrier #oss #foss #julialang #unosolver #python #cpp
Release v2.9.0 · cvanaret/Uno
Uno v2.9.0 implements substantial improvements to the barrier method.
As of Sept 9, 2026, Uno can solve 45/47 instances on the Mittelmann benchmark, in about the same time as IPOPT.
Features
Secon...
github.com
September 10, 2026 at 8:57 PM
𝐔𝐧𝐨 is catching up to the big names on the 𝐌𝐢𝐭𝐭𝐞𝐥𝐦𝐚𝐧𝐧 𝐛𝐞𝐧𝐜𝐡𝐦𝐚𝐫𝐤! ⚡
Uno v2.9.0 stepped up its barrier method and now solves 45/47 𝐢𝐧𝐬𝐭𝐚𝐧𝐜𝐞𝐬
github.com/cvanaret/Uno/releases/tag/v2.9.0
#optimization #mathematicalprogramming #numericaloptimization #barrier #oss #foss #julialang #unosolver #python #cpp
Uno v2.9.0 stepped up its barrier method and now solves 45/47 𝐢𝐧𝐬𝐭𝐚𝐧𝐜𝐞𝐬
github.com/cvanaret/Uno/releases/tag/v2.9.0
#optimization #mathematicalprogramming #numericaloptimization #barrier #oss #foss #julialang #unosolver #python #cpp
the problem with this idea is that im not smart enough to ask claude how exactly to do it
@mraginsky.bsky.social can you get ipopt to use its filter/regularization step to mimic poyntragin's maximum principle and thereby have it "understand" optimal control?
@mraginsky.bsky.social can you get ipopt to use its filter/regularization step to mimic poyntragin's maximum principle and thereby have it "understand" optimal control?
July 24, 2026 at 6:06 AM
the problem with this idea is that im not smart enough to ask claude how exactly to do it
@mraginsky.bsky.social can you get ipopt to use its filter/regularization step to mimic poyntragin's maximum principle and thereby have it "understand" optimal control?
@mraginsky.bsky.social can you get ipopt to use its filter/regularization step to mimic poyntragin's maximum principle and thereby have it "understand" optimal control?
Impatiently watching IPOPT iterate at work is a vocation, but continuing to do so at home is more of a calling.
#biomechsky #compneurosky
#biomechsky #compneurosky
November 3, 2023 at 12:17 AM
Impatiently watching IPOPT iterate at work is a vocation, but continuing to do so at home is more of a calling.
#biomechsky #compneurosky
#biomechsky #compneurosky
If you're like me (obstinate Mac user who likes computer simulation), IPOPT iterates about 20% faster on my new 2026 MacBook Pro (M5 Pro CPU) vs. previous 2024 MacBook Pro (M4 Pro CPU).
Only problem is the charging brick isn't big enough and it drops battery when running Moco, even when plugged in!
Only problem is the charging brick isn't big enough and it drops battery when running Moco, even when plugged in!
August 28, 2026 at 6:26 PM
If you're like me (obstinate Mac user who likes computer simulation), IPOPT iterates about 20% faster on my new 2026 MacBook Pro (M5 Pro CPU) vs. previous 2024 MacBook Pro (M4 Pro CPU).
Only problem is the charging brick isn't big enough and it drops battery when running Moco, even when plugged in!
Only problem is the charging brick isn't big enough and it drops battery when running Moco, even when plugged in!
📄 HyAL Optimizer: New Hybrid Method Beats IPOPT and CMA-ES on Complex Constraints
Evolutionary algorithm plus Lagrangian framework solves high-dimensional constrained optimization...
https://theneuralfeed.com/share/post/H7oWQd9M
#AIResearch #MachineLearning #DeepLearning
Read the full story →
Evolutionary algorithm plus Lagrangian framework solves high-dimensional constrained optimization...
https://theneuralfeed.com/share/post/H7oWQd9M
#AIResearch #MachineLearning #DeepLearning
Read the full story →
theneuralfeed.com
July 21, 2026 at 8:17 AM
📄 HyAL Optimizer: New Hybrid Method Beats IPOPT and CMA-ES on Complex Constraints
Evolutionary algorithm plus Lagrangian framework solves high-dimensional constrained optimization...
https://theneuralfeed.com/share/post/H7oWQd9M
#AIResearch #MachineLearning #DeepLearning
Read the full story →
Evolutionary algorithm plus Lagrangian framework solves high-dimensional constrained optimization...
https://theneuralfeed.com/share/post/H7oWQd9M
#AIResearch #MachineLearning #DeepLearning
Read the full story →
Pyomo+IPOPTだと相補性条件を扱えなくて、書き下すかPATH使えばいいんだけどPyomo側でなんとかならんかと思ってmpecモジュール眺めてたらSolverFactory("mpec_nlp")を使えばいいらしい
February 25, 2025 at 9:21 PM
Pyomo+IPOPTだと相補性条件を扱えなくて、書き下すかPATH使えばいいんだけどPyomo側でなんとかならんかと思ってmpecモジュール眺めてたらSolverFactory("mpec_nlp")を使えばいいらしい
New on CRAN: Uno (2.7.3). View at https://CRAN.R-project.org/package=Uno
Uno: R Interface to the 'Uno' Nonlinear Optimization Solver
Bindings to 'Uno' (Unifying Nonlinear Optimization), a C++ solver for smooth nonlinearly constrained optimization. 'Uno' unifies Lagrange-Newton methods, including sequential quadratic programming and interior-point methods, by decomposing them into interacting building blocks (constraint-relaxation, inequality-handling, Hessian, and globalization strategies) that can be freely combined, either through options or through presets that reproduce established solvers such as 'filterSQP' and 'IPOPT'. The framework is described in Vanaret and Leyffer (2024) <<a href="https://doi.org/10.48550%2FarXiv.2406.13454" target="_top">doi:10.48550/arXiv.2406.13454</a>>.
CRAN.R-project.org
June 8, 2026 at 10:22 PM
New on CRAN: Uno (2.7.3). View at https://CRAN.R-project.org/package=Uno
both drift trajectory tracking and computational efficiency. Our approach resulted in a 38$\%$ reduction in RMSE lateral error and achieved an average computation time that is 75$\%$ lower than that of the Interior Point OPTimizer (IPOPT). [5/5 of https://arxiv.org/abs/2503.11083v1]
March 17, 2025 at 6:00 AM
both drift trajectory tracking and computational efficiency. Our approach resulted in a 38$\%$ reduction in RMSE lateral error and achieved an average computation time that is 75$\%$ lower than that of the Interior Point OPTimizer (IPOPT). [5/5 of https://arxiv.org/abs/2503.11083v1]
constraints. The $\ell_1$-Exact Penalty-Barrier enhanced \texttt{IPOPT} improves performance and robustness by introducing additional inequality constraints and decision variables. However, this comes at the cost of increased computational complexity [2/5 of https://arxiv.org/abs/2503.21502v1]
March 28, 2025 at 6:03 AM
constraints. The $\ell_1$-Exact Penalty-Barrier enhanced \texttt{IPOPT} improves performance and robustness by introducing additional inequality constraints and decision variables. However, this comes at the cost of increased computational complexity [2/5 of https://arxiv.org/abs/2503.21502v1]
We show that FSNet performs remarkably well across a wide range of problem classes, including QP, QCQP, SOCP, and AC Optimal Power Flow, achieving orders-of-magnitude speedups.
Surprisingly, in several nonconvex problems, FSNet even finds better local solutions than IPOPT!
Surprisingly, in several nonconvex problems, FSNet even finds better local solutions than IPOPT!
November 26, 2025 at 6:25 PM
We show that FSNet performs remarkably well across a wide range of problem classes, including QP, QCQP, SOCP, and AC Optimal Power Flow, achieving orders-of-magnitude speedups.
Surprisingly, in several nonconvex problems, FSNet even finds better local solutions than IPOPT!
Surprisingly, in several nonconvex problems, FSNet even finds better local solutions than IPOPT!
discretized using a high-order implicit Runge-Kutta method with flipped Legendre-Gauss-Radau points, resulting in a large-scale nonlinear program (NLP) efficiently solved by state-of-the-art NLP solvers such as Ipopt. This formulation enables [2/5 of https://arxiv.org/abs/2505.03552v1]
May 7, 2025 at 6:06 AM
discretized using a high-order implicit Runge-Kutta method with flipped Legendre-Gauss-Radau points, resulting in a large-scale nonlinear program (NLP) efficiently solved by state-of-the-art NLP solvers such as Ipopt. This formulation enables [2/5 of https://arxiv.org/abs/2505.03552v1]
benchmark tasks including quadratic programming (QP), nonconvex programming, and large-scale AC optimal power flow problems. The results demonstrate that compared to existing solvers (e.g., \texttt{OSQP}, \texttt{IPOPT}) and the latest learning-based [4/5 of https://arxiv.org/abs/2505.04037v1]
May 8, 2025 at 6:05 AM
benchmark tasks including quadratic programming (QP), nonconvex programming, and large-scale AC optimal power flow problems. The results demonstrate that compared to existing solvers (e.g., \texttt{OSQP}, \texttt{IPOPT}) and the latest learning-based [4/5 of https://arxiv.org/abs/2505.04037v1]