A future home for research notes and technical writing.https://coolmoon327.github.io/enas-folio (Astro)https://coolmoon327.github.io/blog/vlan-trunking-for-broadband-and-iptv/https://coolmoon327.github.io/blog/vlan-trunking-for-broadband-and-iptv/A reusable VLAN trunking pattern for carrying broadband and IPTV over one physical link while keeping the traffic domains explicit.Thu, 04 Jan 2024 00:00:00 GMTvlaniptvnetwork-designopenwrtSystems EngineeringYuhang Shenhttps://coolmoon327.github.io/blog/growing-linux-storage-safely/https://coolmoon327.github.io/blog/growing-linux-storage-safely/A safety-first mental model and procedure for extending Linux storage across partitions, LVM layers, and filesystems.Wed, 30 Nov 2022 00:00:00 GMTlinux-storagelvmfilesystemsoperationsSystems EngineeringYuhang Shenhttps://coolmoon327.github.io/blog/reproducible-academic-writing-with-vscode-latex-bibtex-git/https://coolmoon327.github.io/blog/reproducible-academic-writing-with-vscode-latex-bibtex-git/A reproducible academic-writing workflow that joins LaTeX structure, BibTeX references, VS Code tooling, and disciplined Git history.Fri, 28 Oct 2022 00:00:00 GMTlatexvscodebibtexgitreproducible-researchResearch PracticeYuhang Shenhttps://coolmoon327.github.io/blog/engineering-reinforcement-learning-for-edge-systems/https://coolmoon327.github.io/blog/engineering-reinforcement-learning-for-edge-systems/An engineering synthesis of state representations, large discrete action selection, and generative latent models for edge-system control.Mon, 24 Oct 2022 00:00:00 GMTreinforcement-learningrepresentation-learninglarge-action-spacesedge-computingvariational-autoencoderCloud and Edge SystemsYuhang Shenhttps://coolmoon327.github.io/blog/qos-aware-service-composition-in-geo-distributed-clouds/https://coolmoon327.github.io/blog/qos-aware-service-composition-in-geo-distributed-clouds/A clear formulation of QoS-aware service composition and the role of genetic search in geo-distributed cloud workflows.Mon, 24 Oct 2022 00:00:00 GMTservice-compositionquality-of-servicegeo-distributed-cloudopenraasCloud and Edge SystemsYuhang Shenhttps://coolmoon327.github.io/blog/how-communication-networks-move-data/https://coolmoon327.github.io/blog/how-communication-networks-move-data/An end-to-end explanation of protocol layering, reliable transport, congestion control, and routing decisions in communication networks.Sun, 23 Oct 2022 00:00:00 GMTcomputer-networksprotocol-layerscongestion-controlroutingWireless and NetworksYuhang Shenhttps://coolmoon327.github.io/blog/information-theory-for-wireless-systems/https://coolmoon327.github.io/blog/information-theory-for-wireless-systems/A structured path from entropy and mutual information to channel capacity and practical coding ideas for wireless systems.Sun, 23 Oct 2022 00:00:00 GMTinformation-theorychannel-capacitycoding-theorywireless-communicationsWireless and NetworksYuhang Shenhttps://coolmoon327.github.io/blog/multi-agent-reinforcement-learning-in-practice/https://coolmoon327.github.io/blog/multi-agent-reinforcement-learning-in-practice/A practical view of multi-agent reinforcement learning through CTDE, MAPPO, coordination challenges, and implementation choices.Sun, 23 Oct 2022 00:00:00 GMTmulti-agent-reinforcement-learningctdemapposystem-designReinforcement LearningYuhang Shenhttps://coolmoon327.github.io/blog/optimization-for-wireless-systems/https://coolmoon327.github.io/blog/optimization-for-wireless-systems/A problem-driven guide to convexity, Lagrangian duality, KKT conditions, and the optimization patterns used in wireless research.Sun, 23 Oct 2022 00:00:00 GMTconvex-optimizationlagrange-dualitykkt-conditionswireless-systemsWireless and NetworksYuhang Shenhttps://coolmoon327.github.io/blog/reinforcement-learning-from-bellman-to-actor-critic/https://coolmoon327.github.io/blog/reinforcement-learning-from-bellman-to-actor-critic/A unified route from Markov decision processes and Bellman equations to value learning, policy gradients, and actor–critic methods.Sun, 23 Oct 2022 00:00:00 GMTreinforcement-learningbellman-equationactor-criticpolicy-optimizationReinforcement LearningYuhang Shenhttps://coolmoon327.github.io/blog/researcher-macos-workstation/https://coolmoon327.github.io/blog/researcher-macos-workstation/A curated macOS workstation for research, writing, and development, with emphasis on durable workflows rather than app collecting.Sun, 23 Oct 2022 00:00:00 GMTmacosresearch-workflowdeveloper-toolsautomationResearch PracticeYuhang Shenhttps://coolmoon327.github.io/blog/safe-reinforcement-learning-beyond-reward-penalties/https://coolmoon327.github.io/blog/safe-reinforcement-learning-beyond-reward-penalties/A comparison of constrained MDPs, trust-region policy updates, and action-projection safety layers for enforcing explicit constraints.Sun, 23 Oct 2022 00:00:00 GMTsafe-reinforcement-learningconstrained-mdptrust-regionsafety-layerReinforcement LearningYuhang Shenhttps://coolmoon327.github.io/blog/stochastic-models-for-wireless-random-access/https://coolmoon327.github.io/blog/stochastic-models-for-wireless-random-access/A connected treatment of Poisson and MMPP traffic, bursty arrivals, and ALOHA throughput models for wireless random access.Sun, 23 Oct 2022 00:00:00 GMTstochastic-processestraffic-modelingmmppaloharandom-accessWireless and NetworksYuhang Shen