Welcome to EasyCO’s documentation!¶

EasyCO is a learning-driven platform for solving Combinatorial Optimization (CO) problems. We aim to provide the CO community with solvers that are easy-to-use, flexible, and broadly applicable.

pipline:

🚀Features¶

Extensive Coverage: EasyCO supports the solution generation for 55 common CO problems. This broad coverage is complemented by a comprehensive suite of benchmark datasets to facilitate fair and robust comparisons. To the best of our knowledge, EasyCO is the first platform to achieve coverage of over 50 problems.
Flexible Composition: EasyCO is built upon a unified optimization pipeline, where each neural solver is decomposed into interchangeable modules. This modular framework allows the flexible composition of components from different solvers, thereby facilitating the rapid development of novel and tailored solutions.
Intuitive GUI: We provide a user-friendly Graphical User Interface (GUI) that enables a full code-free workflow from experimental design to result analysis. This significantly lowers the technical barrier and accelerates the research process.

🌟Supported tasks¶

**Supported Tasks**¶
Problem Category	Description	Typical Problems
Routing Problems	Find one or more paths in a graph to minimize cost or maximize reward under constraints.	TSP, ATSP, PCTSP, CVRP, OP, SOP
Scheduling Problems	Assign start times or order to tasks under resource/time constraints to optimize goals.	FFSP, RCPSP, SMTWTP
Packing Problems	Select items under capacity limits to maximize value or minimize container usage.	KP, MKP, BPP
Other Graph Problems	Match items or select subsets under structural constraints to optimize objective.	MIS

**Solvers**¶
Problem	Solvers
TSP
CVRP
OP
ATSP
PCTSP
SOP

🎉Supported methods¶

EasyCO supports the following classical methods:

Supported Classical Methods¶
Method	Paper	code
LKH-3	An effective implementation of the Lin-Kernighan traveling salesman heuristic (European Journal of Operational Research, 2000)	http://akira.ruc.dk/~keld/research/LKH-3/
EAX	A Powerful Genetic Algorithm Using Edge Assembly Crossover for the Traveling Salesman Problem (INFORMS Journal on Computing, 2013)	https://github.com/nagata-yuichi/GA-EAX
HGS	Hybrid genetic search for the CVRP: Open-source implementation and SWAP* neighborhood (Computers & Operations Research, 2022)	https://github.com/vidalt/HGS-CVRP
OR-Tools	https://ai.googleblog.com/2019/09/or-tools-now-supports-integer.html
Concorde	Concorde TSP solver (2006)	https://github.com/jvkersch/pyconcorde
PyVRP

EasyCO supports the following NCO methods:

Supported NCO Methods¶
Method	Paper
AM	Attention, Learn To Solve Routing Problems! (ICLR, 2019)
DACT	Learning to Iteratively Solve Routing Problems with Dual-Aspect Collaborative Transformer (NeurIPS, 2021)
DeepACO	DeepACO: Neural-enhanced Ant Systems for Combinatorial Optimization (NeurIPS, 2023)
DIFUSCO	Difusco: Graph-based diffusion solvers for combinatorial optimization (NeurIPS 2023)
DPN	DPN: decoupling partition and navigation for neural solvers of min-max vehicle routing problems (ICML, 2024)
ELG	Towards generalizable neural solvers for vehicle routing problems via ensemble with transferrable local policy (IJCAI, 2024)
GLOP
H-TSP	H-TSP: Hierarchically Solving the Large-Scale Traveling Salesman Problem (AAAI 2023)
ICAM	Instance-Conditioned Adaptation for Large-scale Generalization of Neural Combinatorial Optimization (2024)
INViT	Invit: A generalizable routing problem solver with invariant nested view transformer (ICML, 2024)
L2S	Deep Reinforcement Learning Guided Improvement Heuristic for Job Shop Scheduling (ICLR 2024)
LEHD	Neural combinatorial optimization with heavy decoder: Toward large scale generalization (NeurIPS, 2023)
LIH	Learning Improvement Heuristics for Solving Routing Problems (TNNLS, 2021)
MatNet	Matrix Encoding Networks for Neural Combinatorial Optimization (NeurIPS, 2021)
NLNS	Neural large neighborhood search for routing problems (Artificial Intelligence, 2022)
OMNI	Towards Omni-generalizable Neural Methods for Vehicle Routing Problems (ICML, 2023)
POMO	POMO: Policy Optimization with Multiple Optima for Reinforcement Learning (NeurIPS, 2020)
T2T	T2t: From distribution learning in training to gradient search in testing for combinatorial optimization (NeurIPS, 2023)
UDC	UDC: a unified neural divide-and-conquer framework for large-scale combinatorial optimization problem (NeurIPS, 2024)

Note

1111

Tip

2222

🤝About EasyNCO¶

📄Citation¶

@inproceedings{ye2023deepaco,
  title={DeepACO: Neural-enhanced Ant Systems for Combinatorial Optimization},
  author={Ye, Haoran and Wang, Jiarui and Cao, Zhiguang and Liang, Helan and Li, Yong},
  booktitle={Advances in Neural Information Processing Systems},
  year={2023}
}