Dennis O.

STL AddOns™ -- it started with a hypothesis: Could we match freight to partially loaded trucks that still have unused trailer space? We'd have to do this dynamically, in real time. Which would mean developing novel algorithms that can continually process through trillions of shipment combinations in seconds. And developing ML models to predict cost curves, pooling probability, dynamic market conditions, and more. And then integrating these with our existing Shared Truckload platform to create a simple user experience for dispatchers and drivers. That's how we eliminate unused trailer space so carriers can turn air into earnings! To learn more about Shared Truckload, check out our latest article (link in the comments).

55

2 Comments

To my academic followers (and anyone else who took a course in optimization):   The optimization community has long taught that:   ·       Linear programming, popularized by Dantzig, is a powerful tool, widely used for many applications. ·       Dynamic programming, pioneered by Bellman, is elegant theory, primarily for Ph.D. students and with limited applicability.   Last April I gave a talk at Cornell’s ORIE department that challenged both of these points. By the end of the talk, I made the points:   ·       Linear programming is useful, but only on a narrow subset of decision problems (and most of these are sequential). ·       Sequential decision problems are virtually universal, something we all encounter every day… but Bellman’s equation is rarely used, limited to a small number of applications.   We can dramatically expand the usefulness of what we teach in an optimization course. My talk outlines how to do this. I have broken the talk into four parts on the webpage:   https://lnkd.in/e_kim9u9 “tinyurl.com/” with “HowtoTeachOptimization”

1,310

50 Comments

Orchestration Software - so what? Like many other topics, most of us have become desensitized to the term "Orchestration Software" primarily because of the high-level hype. My hope is that this, my latest paper of a series, provides enough details and granularity that you can better slice through the hype and examine the granular details for yourself, so that you can accurately evaluate the requirements of a true orchestration software stack. In this paper I discuss: *Software roles *The importance of real-time decision making and key functionality *Safeguarding throughput *Exception handling *Simulation and modeling *The evaluation criteria of Orchestration Software In today's world of automated warehouse solutions, mechanical speeds and rates make throughput theoretically possible. But it's the orchestration software that makes throughput attainable in real world operations. "Having conversations that aren't being had about warehouse automation." #warehouseautomation #orchestration #WES #orchestrationsoftware

9

5 Comments