JuMP Developers Meetup/Workshop

June 12-16, 2017, at Sloan School of Business, Massachusetts Institute of Technology.

The workshop was sponsored by the MIT Sloan Latin America Office.

Purpose

The workshop is designed to serve as an opportunity for developers of mathematical optimization software within the JuMP “stack” (i.e., solvers, solver interfaces, MathProgBase, JuMP, and JuMP extensions) to meet and focus on advancing common interests. JuMP core developers will present on the internals of JuMP and solicit feedback in preparation for the release of JuMP 1.0 (targeted for July, 2017). Attendees are welcome to contribute to the JuMP stack itself or to pursue JuMP-related projects while having the JuMP developers in the same room for questions and feedback.

Locations

On Monday morning until the lunch break, Tuesday all day, and Wednesday all day, the workshop will take place in the room E62-223. This is on the second floor of the main building of MIT Sloan.

Thursday all day and Friday all day, the workshop will take place in the room E62-650. This is on the sixth floow of the main building of MIT Sloan.

Schedule

Monday June 12

Room E62-223 (until lunch)

09:00 Welcome and Introduction (Juan Pablo Vielma, MIT) [Slides] [Video]

09:25 Overview and Welcome (Miles Lubin, MIT) [Video]

09:30 The design of JuMP and MathProgBase (Miles Lubin, MIT) [Slides] [Video]

10:30 The design and architecture of Pajarito (Chris Coey, MIT) [Slides] [Video]

11:00 Coffee break and group photo

11:45 A Graph-Based Architecture for Optimization Modeling Frameworks (Steven Diamond, Stanford) [Video]

12:15 Lunch break

Room E51-376 (after lunch)

14:30 Plans and priorities for JuMP 1.0 and MathProgBase 1.0 (Miles Lubin)

15:00 Discussion and feedback on JuMP 1.0 and MathProgBase 1.0, discuss important issues including MathProgBase issue #164 on solver statuses and MathProgBase issue #168 on a new standard problem format

17:00 Finish

Tuesday June 13

Room E62-223 (all day)

09:10 Welcome for the day

09:15 Sum-of-squares optimization in Julia [remote presentation] (Benoît Legat, Université Catholique de Louvain) [Slides] [Video]

09:45 PowerModels.jl: A Brief Introduction (Carleton Coffrin, Los Alamos National Lab) [Slides] [Video]

10:15 Coffee break

10:45 Stochastic programming in energy systems (Joaquim Dias Garcia, PSR and PUC-Rio) [Slides] [Video]

11:15 Graph-based modeling using JuMP (Jordan Jalving, UW-Madison) [Slides] [Video]

11:45 Decarbonization of power systems using JuMP (Nestor Sepulveda, MIT)

12:15 Lunch break

14:00 Working with JuMP’s macros for extensions (Yee Sian Ng, MIT) [Notebook] [Video]

14:30 PiecewiseLinearOpt.jl: Solving optimization problems containing piecewise linear functions (Joey Huchette, MIT) [Slides] [Video]

15:00 OSQP.jl: A Julia wrapper and extensions for the Operator Splitting QP solver (Bartolomeo Stellato, Oxford) [Slides] [Video]

15:30 SolverStudio + Julia (Oscar Dowson, University of Auckland) [Video]

16:00 Finish

Wednesday June 14

Room E62-223 (all day)

09:00 Tutorial and live coding with Documenter.jl

11:00 A talk on automatic differentiation using ForwardDiff.jl and ReverseDiff.jl (Jarrett Revels, MIT) [Video]

11:45 Lunch break

14:00 Developer collaboration

17:00 Finish

Thursday June 15

Room E62-650 (all day) Updated

09:00 Developer collaboration, brainstorming on JuMP’s containers (JuMP issue #1047)

12:00 Lunch break

14:00 Developer collaboration, brainstorming on solver callbacks (MathProgBase issue #170)

16:00 Finish

18:00 Workshop dinner

Friday June 16

10:00 Tour of the MIT Museum (meet at 265 Massachusetts Avenue, Cambridge)

11:30 Lunch break

Room E62-650 Updated

14:00 Developer collaboration, brainstorming on the future of algebraic modeling languages (Chris Coey)

17:00 Finish

Contact

Contact mlubin at mit.edu for more information.