arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2611.06879cs.LGcs.NEmath.OC
收藏

自适应多模式:基于在线景观探测的预算感知优化器路由

Adaptive Multi Mode: Budget-Aware Routing of Optimizers By Online Landscape Probes

Benjamin Jacob Bodner

AI总结:

本研究提出 AMM(自适应多模式),通过在线探测优化景观特征在多种优化引擎间路由预算,在 bbob 等基准测试中多数场景下优于竞品,已发布为开源 Python 包。

摘要:

针对连续黑箱问题,最优经典优化器取决于其预算和问题景观。现有路由方法仅依据维度、预算、过往改进或适配单一基准的特征决策,极少在线测试优化器的隐式假设。AMM(自适应多模式)通过在线探测曲率、平台区、可分性、扭结、 basin( basin 指优化景观中的吸引盆)及分形结构,在拟牛顿法、信赖域法、坐标法、多起点法和进化策略优化引擎间分配已知预算。每个探测有默认设置、按维度-预算划分的有效范围及从预算中扣除的成本。在 bbob 基准的预注册开发循环中,最后 16 次确认的 36 项预测中有 32 项成立。随后,AMM 在预留的 bbob 实例 101-110(维度 D=2-40,预算范围 10*D-1000*D)上与最多 7 种竞品进行单次测试:在 600 个函数-维度-预算单元中,AMM 在 382 个单元中具有最高目标占比(236 个单元独占最高,146 个单元并列最高);依据描述性后验规则,AMM 在 35 个单元中明显落后,且主要集中在预算 10*D 和 30*D 处。在各预算下,AMM 均领先于由其自身引擎构建的对照组(单独最优引擎、按维度和预算的最优引擎、贪心调度),在 10*D 和 30*D 处领先幅度较小,同时也领先于 nevergrad 的 NgIohTuned。在预注册运行中,AMM 对 IPOP-CMA-ES 的领先在仿射变换和盒约束变换下均成立,在盒约束 bbob 基准及合并后的旋转 f2-f4 基准上也成立,但旋转可分的 f3 和 f4 基准会使该领先大幅降低。在预留的 MA-BBOB 基准上,AMM 在预算 2000*D 处的最终占比与 IPOP-CMA-ES 持平,但其任意时刻性能面积更低。我们已将 AMM 作为开源 Python 包发布(pip install amm-optimizer;此 https URL)。

英文摘要:

The best classical optimizer for a continuous black-box problem depends on its budget and landscape. Existing routers decide from the dimension and budget, past improvement or features fitted to one benchmark, and few test an optimizer's implicit assumptions online. AMM (Adaptive Multi Mode) routes a known budget among quasi-Newton, trust-region, coordinate, multistart and evolution-strategy optimization engines with online probes of curvature, plateaus, separability, kinks, basins and fractal structure. Each probe has a default, a dimension-by-budget validity range and a cost charged to the budget. In its pre-registered development loop on bbob, 32 of 36 predictions held over the last 16 confirmations. AMM was then tested once on held-out bbob instances 101-110 (D=2-40, 10*D-1000*D) against up to seven competitors. It has the highest target share in 382 of 600 function-dimension-budget cells (236 outright, 146 tied) and, by a descriptive posterior rule, is credibly behind in 35, mostly at 10*D and 30*D. At every budget it also leads the controls built from its own engines (the best engine alone, the best engine per dimension and budget, and a greedy schedule), narrowly at 10*D and 30*D, and nevergrad's NgIohTuned. In pre-registered runs its lead over IPOP-CMA-ES holds under affine and box transformations, on box-constrained bbob and, pooled, on rotated f2-f4, but rotating the separable f3 and f4 removes most of the lead there. On held-out MA-BBOB its final share at 2000*D is level with IPOP-CMA-ES's and its anytime area lower. We published AMM as an open-source Python package (pip install amm-optimizer; https://github.com/BjBodner/amm-optimizer).

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑