AGENT SKILL

cortexcloud-formulate

暂无简介

License 许可
Apache-2.0
最近更新
2026-10-24
Stars
★ 0
Forks
0

02 资源文件 1

03 SKILL.md 正文

cortexcloud-formulate

CortexCloud Formulate

Domain-to-QUBO adapter library for the CortexCloud optimization API. One function per problem domain — takes real-world objects (returns, distances, shifts, blocks) and emits QUBO matrices that can be sent directly to POST https://api.cortexcloud.org/v1/optimize.

Repo: https://github.com/jonahthan433/cortexcloud-formulate Install: Copy cortexcloud_formulate.py + pip install numpy. Single file, no framework dependencies.

Domains Covered

Function Problem Input
portfolio() Cardinality-constrained Markowitz Returns, covariance, K
tsp() Travelling salesman (time-indexed) Distance matrix
assignment() Task/worker or SKU/warehouse Cost matrix
scheduling() Single-machine scheduling Processing times, deadlines
nurse_scheduling() Nurse-to-shift preference allocation Preference matrix
mining_cutoff() Open-pit block extraction Grades, precedence edges
unit_commitment() Generator dispatch Profits matrix, startup costs
maxcut() Graph MAX-CUT Adjacency matrix
crop_allocation() Field crop allocation Yields, adjacency edges

Usage Pattern

from cortexcloud_formulate import portfolio, solve
import numpy as np

returns = np.array([0.05, -0.02, 0.03, -0.04, 0.01, 0.02])
cov = np.diag([0.01, 0.02, 0.015, 0.025, 0.01, 0.02])
problem = portfolio(returns, cov, cardinality=3)

result = solve(problem)
print(result["recommendation"]["solver_id"])   # "brute-force"
print(result["recommendation"]["cortexcloud_price_usd"])  # 0.05

Key Design Decisions

  • Only QUBO encoding, no solver logic. CortexCloud handles solver selection, routing, and execution.
  • Constraint penalties auto-scaled proportionally to objective values.
  • n = variables² for time-indexed formulations (TSP, scheduling). Keep small for exact solvers.
  • All formulations minimise — profits are negated, costs stay positive.

Testing

cd cortexcloud-formulate && source .venv/bin/activate
python -c "from cortexcloud_formulate import *; print('ok')"
python tests/demo.py

Venv

Local .venv (Python 3.12, numpy 2.5.2) via uv venv && uv pip install numpy.

See Also

  • cortexcloud-platform skill — API architecture, deploy, router bugs
  • references/benchmarks-trial-endpoints.md in cortexcloud-platform

04 同仓库更多 Skills