CP 2026 Workshop on
Soft Constraints, Discrete Optimization, and Machine Learning (SOFT’2026)
Lisbon, Portugal
July 24, 2026 (Friday)
As part of FloC 2026
Motivation and Scope
Since Freuder’s seminal work on Partial Max-CSP in 1991, research on soft constraints has grown in several directions in the constraint programming and related fields, including Max-SAT, Max-SMT, Markov Random Fields, and pseudo-Boolean optimization.
Today, the convergence of discrete optimization and machine learning is emerging as a game-changing force in the realm of AI, opening new perspectives for constraint reasoning, optimization, and learning. This workshop aims to bring together researchers from these communities to present recent advances, share ongoing work, and discuss future directions for hybrid approaches that combine discrete optimization with machine learning and data mining.
Format
The workshop will feature:
● Two invited talks highlighting recent results (some already published).
● Contributed talks selected from extended abstracts or (short) papers describing ongoing work.
Topics of Interest (but are not restricted to):
- ● Max-SAT
- ● Max-SMT
- ● Markov Random Field
- ● Pseudo-Boolean Optimization
- ● Soft global constraints
- ● Weighted CSP
- ● Integer Programming
- ● Combining discrete optimization with machine learning for better solver design.
- ● Data-driven strategies to guide search heuristics, branching, or propagation.
- ● Using machine learning and data mining techniques to guide search.
- ● Integrating deep neural networks to improve solvers.