<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
  <title>OpEra Stories: operations research in industry</title>
  <link>https://operatechnologies.com/stories</link>
  <atom:link href="https://operatechnologies.com/rss.xml" rel="self" type="application/rss+xml" />
  <description>Documented cases where a model took over a planning decision, with the original papers.</description>
  <language>en</language>
  <lastBuildDate>Thu, 22 Oct 2026 12:00:00 GMT</lastBuildDate>
  <item>
    <title>Molslinjen forecast every ferry departure a year ahead and estimated up to $3.2M a year in extra profit</title>
    <link>https://operatechnologies.com/stories/molslinjen-ferry-forecasting</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/molslinjen-ferry-forecasting</guid>
    <pubDate>Thu, 22 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Danish ferry line Molslinjen forecasts each departure up to a year ahead to set fares and pack car decks: an estimated 15–20M DKK extra profit a year.</description>
    <category>Ferry revenue management</category>
  </item>
  <item>
    <title>Chevron has planned its seven refineries with one linear program for 30 years and values it near $1B a year</title>
    <link>https://operatechnologies.com/stories/chevron-refinery-planning</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/chevron-refinery-planning</guid>
    <pubDate>Wed, 21 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Chevron plans crudes, products and unit operation at seven refineries with one linear program: near $1B of value a year, by its own estimate.</description>
    <category>Refinery planning</category>
  </item>
  <item>
    <title>Meituan cut delivery time per order by 21% with a model that assigns a whole city's orders every 30 seconds</title>
    <link>https://operatechnologies.com/stories/meituan-courier-dispatch</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/meituan-courier-dispatch</guid>
    <pubDate>Tue, 20 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Meituan assigns 60M+ daily orders to couriers citywide every 30 seconds: delivery time per order down 21% and distance down 24%, 2022 vs 2019.</description>
    <category>Courier dispatch</category>
  </item>
  <item>
    <title>Air New Zealand replaced two weeks of hand-built rosters with an hour of computing, saving an estimated NZ$15.7M a year</title>
    <link>https://operatechnologies.com/stories/air-new-zealand-crew-scheduling</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/air-new-zealand-crew-scheduling</guid>
    <pubDate>Mon, 19 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Air New Zealand and the University of Auckland built eight crew planning and rostering systems: about NZ$15.7M saved a year, by the authors' count.</description>
    <category>Crew scheduling</category>
  </item>
  <item>
    <title>Intel planned its products and factories as one problem and credits it with $25.4B since 2009</title>
    <link>https://operatechnologies.com/stories/intel-product-supply</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/intel-product-supply</guid>
    <pubDate>Sun, 18 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Intel links product design to supply planning across 200+ factories: $1.9B more revenue and $1.5B lower costs a year on average, by Intel's count.</description>
    <category>Product supply planning</category>
  </item>
  <item>
    <title>Walmart took 108,000 truck routes off the road in a year by planning routes and loads together</title>
    <link>https://operatechnologies.com/stories/walmart-load-planner</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/walmart-load-planner</guid>
    <pubDate>Sat, 17 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Walmart's Load Planner plans each route and trailer load together: 108,000 routes and 33M miles cut, $91.5M saved in fiscal 2023 by Walmart's count.</description>
    <category>Routing and loading</category>
  </item>
  <item>
    <title>Microsoft moved its cloud hardware planning from spreadsheets to a model and cut cycle times in half</title>
    <link>https://operatechnologies.com/stories/microsoft-cloud-supply-chain</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/microsoft-cloud-supply-chain</guid>
    <pubDate>Fri, 16 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Microsoft plans hardware for 400+ data centers with an optimization model and a language layer: cycle times halved, answers in near real time.</description>
    <category>Cloud hardware fulfillment</category>
  </item>
  <item>
    <title>CSX let a model assign its empty railcars and estimated it saved over $51M a year</title>
    <link>https://operatechnologies.com/stories/csx-empty-railcars</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/csx-empty-railcars</guid>
    <pubDate>Sun, 11 Oct 2026 12:00:00 GMT</pubDate>
    <description>How CSX assigns empty railcars to customer orders in real time: $51M+ saved a year and $1.4B of railcars it never had to buy, by CSX's estimate.</description>
    <category>Empty railcar distribution</category>
  </item>
  <item>
    <title>UPS changed the order of its drivers' stops and expected to save $300M to $400M a year</title>
    <link>https://operatechnologies.com/stories/ups-orion</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/ups-orion</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>How UPS built ORION to sequence the stops of 55,000 drivers: about 100 million fewer miles a year and $300M to $400M in expected annual savings.</description>
    <category>Route sequencing</category>
  </item>
  <item>
    <title>Waste Management took 984 collection routes off the road in nine months</title>
    <link>https://operatechnologies.com/stories/waste-management</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/waste-management</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>WasteRoute re-sequenced every stop in a district and cut 984 collection routes in nine months, about $18M saved in 2003 by WM's estimate. Sources inside.</description>
    <category>Collection routes</category>
  </item>
  <item>
    <title>Coca-Cola Enterprises cut $45M a year by changing how its dispatchers build routes</title>
    <link>https://operatechnologies.com/stories/coca-cola-enterprises</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/coca-cola-enterprises</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Coca-Cola Enterprises put a routing model under 300+ dispatchers and about 10,000 delivery trucks, for a $45M annual cost saving. Sources inside.</description>
    <category>Fleet routing</category>
  </item>
  <item>
    <title>Schneider National tested a driver policy in a model and avoided a $30M-a-year plan</title>
    <link>https://operatechnologies.com/stories/schneider-national</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/schneider-national</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>Schneider and Princeton built a fleet simulator that reproduces how dispatchers decide, then used it to test policies before rollout. Sources inside.</description>
    <category>Fleet policy</category>
  </item>
  <item>
    <title>Zara added an estimated $586M in revenue from stock it already had</title>
    <link>https://operatechnologies.com/stories/zara</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/zara</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Zara and MIT researchers rebuilt the size-by-size allocation of stock to 1,500 stores, for an estimated 3–4% sales lift. Paper and sources inside.</description>
    <category>Store allocation</category>
  </item>
  <item>
    <title>Kroger took $120M of inventory out of 1,950 pharmacies and ran out of stock less often</title>
    <link>https://operatechnologies.com/stories/kroger</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/kroger</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Kroger set reorder levels per drug and per store with simulation-optimization: $120M less inventory, 1.6M fewer out-of-stocks a year.</description>
    <category>Pharmacy inventory</category>
  </item>
  <item>
    <title>Kellogg has planned its cereal plants with one optimization model since 1990</title>
    <link>https://operatechnologies.com/stories/kellogg</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/kellogg</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>The Kellogg Planning System, a weekly linear program across 5 plants and 7 DCs: $4.5M saved in 1995, $35–40M a year projected from one capacity decision.</description>
    <category>Production planning</category>
  </item>
  <item>
    <title>Procter &amp; Gamble redrew which plant makes what and saved over $200M a year</title>
    <link>https://operatechnologies.com/stories/pg-plant-network</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/pg-plant-network</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>How P&amp;G and the University of Cincinnati used integer programming and maps to redesign its North American plant network: almost 20% fewer plants, $200M+ a year.</description>
    <category>Plant network design</category>
  </item>
  <item>
    <title>Taco Bell saved over $53M of labor by scheduling staff in 15-minute slots</title>
    <link>https://operatechnologies.com/stories/taco-bell</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/taco-bell</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>Taco Bell's labor system: 15-minute forecasts, a queue simulation and an integer program for shifts. Over $40M saved in 1993–96 and $53M through 1997.</description>
    <category>Labor scheduling</category>
  </item>
  <item>
    <title>Hartford Hospital's emergency department cut overtime shifts by 85% with a scheduling model</title>
    <link>https://operatechnologies.com/stories/hartford-hospital</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/hartford-hospital</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>How Hartford Hospital and MIT automated emergency department nurse scheduling: 85% fewer overtime shifts, 8–25% better coverage, about $720K a year projected.</description>
    <category>Nurse scheduling</category>
  </item>
  <item>
    <title>A snowstorm cancelled 112 Continental flights. The crews were re-planned in minutes</title>
    <link>https://operatechnologies.com/stories/continental-crew-recovery</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/continental-crew-recovery</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>CrewSolver re-planned Continental's crews after major disruptions in minutes instead of hours, about $40M saved in 2001 by Continental's estimate.</description>
    <category>Crew recovery</category>
  </item>
  <item>
    <title>Google cut a data centre's cooling energy by up to 40% by letting a model pick the settings</title>
    <link>https://operatechnologies.com/stories/google-datacenter-cooling</link>
    <guid isPermaLink="true">https://operatechnologies.com/stories/google-datacenter-cooling</guid>
    <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
    <description>How DeepMind's models first recommended, then directly controlled, Google's data centre cooling: up to 40% less cooling energy in 2016, about 30% in 2018.</description>
    <category>Cooling control</category>
  </item>
</channel>
</rss>
