<?xml version="1.0" encoding="utf-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>供应链 · 拾贰画生 · 有那么点儿追求的中年男人</title><link>https://12lab.cn/categories/%E4%BE%9B%E5%BA%94%E9%93%BE/</link><description>有那么点儿追求的中年男人</description><language>zh-cn</language><atom:link href="https://12lab.cn/categories/%E4%BE%9B%E5%BA%94%E9%93%BE/index.xml" rel="self" type="application/rss+xml"/><item><title>多给十秒，路线就会更好吗：把业务约束与求解参数分开实验</title><link>https://12lab.cn/blog/routing-parameter-experiments/</link><guid isPermaLink="true">https://12lab.cn/blog/routing-parameter-experiments/</guid><pubDate>Tue, 17 Nov 2026 09:00:00 +0800</pubDate><description>沿用容量和时间窗模拟案例，以单因素实验区分业务约束与求解设置，核验路线、等待与求解状态，并保留没有改善的结果。</description></item><item><title>多数日子没有订单，偶尔一单很大：零需求该怎样预测</title><link>https://12lab.cn/topics/forecasting/09-intermittent-demand/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/09-intermittent-demand/</guid><pubDate>Tue, 17 Nov 2026 00:00:00 +0800</pubDate><description>从零值语义开始，手算Croston、SBA和TSB，用两个模拟序列比较间歇需求预测与评价边界。</description></item><item><title>车装得下，路线才成立：容量约束如何改变配送方案</title><link>https://12lab.cn/blog/capacity-constrained-routing/</link><guid isPermaLink="true">https://12lab.cn/blog/capacity-constrained-routing/</guid><pubDate>Sun, 15 Nov 2026 09:00:00 +0800</pubDate><description>沿用十点双车模拟案例，用OR-Tools求解容量约束车辆路径，区分累计服务量与实际载荷，并独立核验覆盖、回仓、容量和距离。</description></item><item><title>总部说卖100件，门店却要110件：预测怎样先对齐，再进入计划</title><link>https://12lab.cn/topics/forecasting/08-hierarchical-reconciliation/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/08-hierarchical-reconciliation/</guid><pubDate>Sun, 15 Nov 2026 00:00:00 +0800</pubDate><description>用匿名服装与禾味模拟案例，逐步手算层级预测协调、整数分配与供给缺口，区分加总一致和预测准确。</description></item><item><title>最短路为什么还不是配送方案：从两点距离到两车基线</title><link>https://12lab.cn/blog/shortest-path-routing-baseline/</link><guid isPermaLink="true">https://12lab.cn/blog/shortest-path-routing-baseline/</guid><pubDate>Sat, 14 Nov 2026 09:00:00 +0800</pubDate><description>用四节点手算和两组十点双车模拟案例，区分最短路、TSP与VRP，建立含回仓距离、客户覆盖与容量检查的可复算配送基线。</description></item><item><title>供应商说平均四天到货，为什么仍会缺货：把交期不确定性放进预测与计划</title><link>https://12lab.cn/topics/forecasting/07-lead-time-uncertainty/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/07-lead-time-uncertainty/</guid><pubDate>Sat, 14 Nov 2026 00:00:00 +0800</pubDate><description>从交期均值与波动到联合情景，用禾味和匿名服装模拟算清等待期间需求、服务指标与未收齐订单。</description></item><item><title>两种补货策略如何权衡库存与缺货</title><link>https://12lab.cn/blog/inventory-test-stockout-tradeoff/</link><guid isPermaLink="true">https://12lab.cn/blog/inventory-test-stockout-tradeoff/</guid><pubDate>Fri, 13 Nov 2026 09:00:00 +0800</pubDate><description>沿用D19冻结配置，在最后28天模拟需求中比较两种补货策略，解释平均库存、满足率、失销与订单次数，并公开逐日流水和复现代码。</description></item><item><title>采购、调拨还是生产：把预测缺口变成可比较的供给计划</title><link>https://12lab.cn/topics/forecasting/06-source-allocation/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/06-source-allocation/</guid><pubDate>Fri, 13 Nov 2026 00:00:00 +0800</pubDate><description>从固定费用反例到双案例完整枚举，用采购、调拨与生产算清资源约束、期望费用与保供边界。</description></item><item><title>库存回放：同一需求下，两种补货策略如何取舍</title><link>https://12lab.cn/blog/inventory-validation-policy-freeze/</link><guid isPermaLink="true">https://12lab.cn/blog/inventory-validation-policy-freeze/</guid><pubDate>Thu, 12 Nov 2026 09:00:00 +0800</pubDate><description>用禾味与匿名服装模拟案例，在28天验证段比较固定批量与周期补货，逐日核验库存守恒、未来信息边界和尾单，再冻结参数供D20测试。</description></item><item><title>预测之后，究竟该订多少：从分位数到受约束的补货建议</title><link>https://12lab.cn/topics/forecasting/05-forecast-to-replenishment/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/05-forecast-to-replenishment/</guid><pubDate>Thu, 12 Nov 2026 00:00:00 +0800</pubDate><description>从保护期、成本分位数和库存位置，到到货、效期、箱规与产能，用双案例算清预测如何成为可审核的建议。</description></item><item><title>安全库存怎么算：先把两种服务水平分开</title><link>https://12lab.cn/blog/safety-stock-service-level/</link><guid isPermaLink="true">https://12lab.cn/blog/safety-stock-service-level/</guid><pubDate>Wed, 11 Nov 2026 09:00:00 +0800</pubDate><description>用禾味与匿名服装模拟案例区分周期无缺货率、单位满足率，手算保护期安全库存，核验正态假设、失销时序与验证冻结边界。</description></item><item><title>从点预测到概率预测与预售取消</title><link>https://12lab.cn/topics/forecasting/04-probability-and-presale/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/04-probability-and-presale/</guid><pubDate>Wed, 11 Nov 2026 00:00:00 +0800</pubDate><description>从需求分位数与概率评分到预售条件留存，用独立模拟实验解释不确定性、风险集与协同边界。</description></item><item><title>到复核日，再补到目标：把周期补货写成可核验规则</title><link>https://12lab.cn/blog/periodic-review-order-up-to/</link><guid isPermaLink="true">https://12lab.cn/blog/periodic-review-order-up-to/</guid><pubDate>Tue, 10 Nov 2026 09:00:00 +0800</pubDate><description>用禾味与匿名服装模拟台账拆解定期复核(R,S)：先判断复核日，再按库存位置补到目标，核验日历、在途和到货时序。</description></item><item><title>从一条直线到多特征预测</title><link>https://12lab.cn/topics/forecasting/03-regression-and-features/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/03-regression-and-features/</guid><pubDate>Tue, 10 Nov 2026 00:00:00 +0800</pubDate><description>从四点手算到多特征与交互项，用禾味两店8SKU模拟数据真正拟合回归，并按时间检查未来误差。</description></item><item><title>预测之前，先把问题与数据说清楚</title><link>https://12lab.cn/topics/forecasting/02-problem-and-data/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/02-problem-and-data/</guid><pubDate>Mon, 09 Nov 2026 15:00:00 +0800</pubDate><description>从模拟食材耗用与服装订单出发，明确预测目标、单位、信息截止点，用14天数据检查主键、缺失与缺货。</description></item><item><title>到阈值就补一批：把固定批量策略写成可核验规则</title><link>https://12lab.cn/blog/fixed-batch-reorder-threshold/</link><guid isPermaLink="true">https://12lab.cn/blog/fixed-batch-reorder-threshold/</guid><pubDate>Mon, 09 Nov 2026 09:00:00 +0800</pubDate><description>沿用禾味与匿名服装七日模拟台账，把日末IP不高于阈值就订固定Q写成规则，核验等号、阈值下穿、每日一单与到货时序。</description></item><item><title>库存够不够，先看时点：库存位置与到货时序</title><link>https://12lab.cn/blog/inventory-position-event-timing/</link><guid isPermaLink="true">https://12lab.cn/blog/inventory-position-event-timing/</guid><pubDate>Sun, 08 Nov 2026 09:00:00 +0800</pubDate><description>用禾味与匿名服装的模拟台账，拆开可用现货、有效在途与欠交，逐日验证到货、需求和下单的先后，避免重复补货与提前使用库存。</description></item><item><title>先做基线，再谈智能：需求预测的模型选型边界</title><link>https://12lab.cn/blog/forecast-model-selection-boundaries/</link><guid isPermaLink="true">https://12lab.cn/blog/forecast-model-selection-boundaries/</guid><pubDate>Sat, 07 Nov 2026 09:00:00 +0800</pubDate><description>汇总禾味与匿名服装的模拟预测证据，说明任务和指标如何改变排名，再给SARIMA、Croston类与树模型划清输入、验证和业务使用边界。</description></item><item><title>数学桥梁：把预测公式算明白</title><link>https://12lab.cn/topics/forecasting/01-math-bridge/</link><guid isPermaLink="true">https://12lab.cn/topics/forecasting/01-math-bridge/</guid><pubDate>Fri, 06 Nov 2026 15:00:00 +0800</pubDate><description>从两个模拟案例出发，逐步读懂预测时点、加权平均、误差、条件概率与logit，附可复现的数学练习包。</description></item><item><title>滚动回测：预测时还不知道的事，不能提前用</title><link>https://12lab.cn/blog/rolling-backtest-stockout-censoring/</link><guid isPermaLink="true">https://12lab.cn/blog/rolling-backtest-stockout-censoring/</guid><pubDate>Fri, 06 Nov 2026 09:00:00 +0800</pubDate><description>用禾味与匿名服装的模拟案例，逐起点回放需求预测，区分一步与固定五步、防止未来信息泄漏，并以独立缺货反例说明销量不等于需求。</description></item><item><title>需求预测怎么比较：平均偏差接近零，为什么仍然不准</title><link>https://12lab.cn/blog/forecast-error-bias-metrics/</link><guid isPermaLink="true">https://12lab.cn/blog/forecast-error-bias-metrics/</guid><pubDate>Thu, 05 Nov 2026 09:00:00 +0800</pubDate><description>用禾味与匿名服装的模拟需求，比较MAE、RMSE、ME和WAPE，拆开工作日与周末，识别误差抵消、零需求和指标排名的边界。</description></item><item><title>简单指数平滑：参数调大，需求预测就更准吗</title><link>https://12lab.cn/blog/simple-exponential-smoothing-demand-forecast/</link><guid isPermaLink="true">https://12lab.cn/blog/simple-exponential-smoothing-demand-forecast/</guid><pubDate>Wed, 04 Nov 2026 09:00:00 +0800</pubDate><description>用禾味和匿名服装的原有模拟需求比较三个平滑参数，分清当日水平与明日预测，并验证单日误差和五天总量为何给出不同答案。</description></item><item><title>移动平均需求预测：最近几天的均值，能代表明天吗</title><link>https://12lab.cn/blog/moving-average-demand-forecast/</link><guid isPermaLink="true">https://12lab.cn/blog/moving-average-demand-forecast/</guid><pubDate>Tue, 03 Nov 2026 09:00:00 +0800</pubDate><description>沿用D9模拟需求，比较3、7、14天移动平均，分清平滑与预测、日误差与五天累计误差，检查未来泄漏和活动后的滞后。</description></item><item><title>季节性朴素预测：用上周同一天建立需求基线</title><link>https://12lab.cn/blog/seasonal-naive-demand-forecast/</link><guid isPermaLink="true">https://12lab.cn/blog/seasonal-naive-demand-forecast/</guid><pubDate>Mon, 02 Nov 2026 09:00:00 +0800</pubDate><description>用固定种子的180天模拟需求，手算七天预测，检查时间切分、未来泄漏与异常周的影响。</description></item><item><title>从OTIF到库存指标 一套可复核的供应链指标字典</title><link>https://12lab.cn/blog/supply-chain-metrics-evidence-package/</link><guid isPermaLink="true">https://12lab.cn/blog/supply-chain-metrics-evidence-package/</guid><pubDate>Sun, 01 Nov 2026 09:00:00 +0800</pubDate><description>用六张指标卡、三项SQL和反例测试，把供应链监控结果接到预测、补货与配送的数据问题上。</description></item><item><title>履约验收通过以后 库存周转、覆盖和损耗该怎样分开算</title><link>https://12lab.cn/blog/inventory-turnover-coverage-loss-validation/</link><guid isPermaLink="true">https://12lab.cn/blog/inventory-turnover-coverage-loss-validation/</guid><pubDate>Sat, 31 Oct 2026 09:00:00 +0800</pubDate><description>用六类履约边界测试、匿名服装生产协同案例和禾味模拟台账，分清历史周转、未来覆盖与损耗的计算边界。</description></item><item><title>三条SQL跑通以后 怎样证明供应链指标算对了</title><link>https://12lab.cn/blog/sql-supply-chain-metrics-boundary-tests/</link><guid isPermaLink="true">https://12lab.cn/blog/sql-supply-chain-metrics-boundary-tests/</guid><pubDate>Fri, 30 Oct 2026 09:00:00 +0800</pubDate><description>用匿名服装与禾味模拟案例，将数量满足率、整单OTIF和快照缺货率写成SQL，并用手算、边界测试和中间结果验证。</description></item><item><title>删掉异常后满足率更高了 把数据问题与业务表现分开</title><link>https://12lab.cn/blog/supply-chain-data-quality-vs-performance/</link><guid isPermaLink="true">https://12lab.cn/blog/supply-chain-data-quality-vs-performance/</guid><pubDate>Thu, 29 Oct 2026 09:00:00 +0800</pubDate><description>用服装与禾味模拟案例，区分重复、缺失、负数和未知库存，核对异常处理怎样改变指标分子、分母与可评价范围。</description></item><item><title>同一个满足率为什么有两个答案 用订单行预聚合修复连接扇出</title><link>https://12lab.cn/blog/sql-join-fanout-preaggregation/</link><guid isPermaLink="true">https://12lab.cn/blog/sql-join-fanout-preaggregation/</guid><pubDate>Wed, 28 Oct 2026 09:00:00 +0800</pubDate><description>订单行连接多条验收记录后，计划量可能被重复累计。用订单行预聚合修复连接扇出，让满足率可复算、可验收。</description></item><item><title>总量收齐了为什么仍不达标 理解数量满足率 OTIF 与快照缺货率</title><link>https://12lab.cn/blog/supply-chain-fulfillment-otif-stockout/</link><guid isPermaLink="true">https://12lab.cn/blog/supply-chain-fulfillment-otif-stockout/</guid><pubDate>Tue, 27 Oct 2026 09:00:00 +0800</pubDate><description>用匿名服装企业与虚构餐饮品牌“禾味”双案例，把“交得怎么样”变成数量满足率、OTIF 与快照缺货率三个可复算指标。</description></item></channel></rss>