Research · 研究
研究Research
这里收集我正在做的研究、长期笔记,以及一些还在成形的项目。
和随笔不同,这一栏更偏向持续、结构化的思考——写得更慢,改得更久,也更愿意把话说清楚。公式、推导和算法都会写出来,不藏在文字后面。
A place for the research I'm doing, long-running notes, and projects still taking shape.
Unlike the essays, this section leans toward sustained, structured thinking — written more slowly, revised for longer, and willing to spell things out. The equations, derivations, and algorithms are on the page, not hidden behind the prose.
研究 — ResearchResearch — 研究
01 / 篇pieces贝尔曼方程：形式化、求解与扩展
The Bellman Equation: Formulation, Solution Methods, and Extensions
从贝尔曼期望方程与最优性方程出发,梳理四类求解方法,以及 TD、Q-learning、DQN、软与值分布形式直至 HJB 方程的衍生脉络。
The Bellman Equation: Formulation, Solution Methods, and Extensions
贝尔曼方程：形式化、求解与扩展
From the Bellman expectation and optimality equations through four families of solution methods, and on to the derived forms — TD, Q-learning, DQN, the soft and distributional variants, and the HJB equation.