Discover and read the best of Twitter Threads about #beigeacademy
Most recents (8)
Machine scheduling, operations research, production, and machine learning. An anecdote thread for @sidwindc.
Biographical background: from age 15 to about 21, I spent my summers working for the laboratory equipment manufacturer that also employed my dad (in "construction").
Biographical background: from age 15 to about 21, I spent my summers working for the laboratory equipment manufacturer that also employed my dad (in "construction").
After that I moved to prepoduction at Siemens, also for healthcare equipment, mostly coding production and tooling processes in CNC. This is where I stayed until my master's thesis.
I might be somewhere in this picture.
I might be somewhere in this picture.
For my master's thesis, and this is the topic of this thread, I moved to what is now Novartis in Basel into the scheduling group. This was both moving from a blue collar to a PhD level white collar environment and from the Bundesliga to the Champions League.
Anecdote time. In the months when I was traveling to visit supply chain crisis managers I was also invited to the ESA mission control center in Darmstadt, ESOC. The mission control room looked exactly as one would expect from the Apollo era: hushed silence and huge monitors.
We were gazing into the control room from a somewhat unkempt anteroom when our tour guide, a true rocket scientist, pointed toward an inconspicuous whiteboard in a corner that looked worse for the wear.
"If there is anything going wrong in here *points at mission control room*, all the experts assemble out here *points at disheveled whiteboard* to work out a solution. And once they agree on a solution, they all sign off here *points at a row of sign-off boxes at the top*."
Little game I played last week: community poker ♠️♥️♣️♦️. It's like poker, but players can trade or pool their cards.
Depending on how well the players know each other, we can get thru all stages of exchange: informal (gift exchange), formal (bargaining & trade), hierarchical.
Depending on how well the players know each other, we can get thru all stages of exchange: informal (gift exchange), formal (bargaining & trade), hierarchical.
It's a 3-minute game that makes the jump from "parlor game" (what everyone else thinks of when they hear "game theory") to strategic interaction, and sets up a very fast emerging exchange economy with wide-ranging compete/cooperate/coordinate ramifications.
The theoretically optimal solution is in most cases to put all cards on one table and to figure out the combination that scores the most points. This usually runs into a variety of practical, often very pedestrian problems.
Lots of confusion still about the very concept of an intermediary (see the many comments below), so time to sort this out.
First, an intermediary is traditionally a provider of a service to facilitate a transaction between original producer and ultimate consumer.
First, an intermediary is traditionally a provider of a service to facilitate a transaction between original producer and ultimate consumer.
Without the intermediary, this transaction would either be very costly or impossible for the end participants to undertake. Bundling less-than-truckload wares into a shipping container is such an intermediary service. Wholesalers are intermediaries.
Indeed marketplaces are intermediaries when they are not operated by the end participants. This has led to quite a bit of confusion among economists who are used to think of a market as an inter-entity abstraction and not as an entity in itself.
Let's create a stylized but very common supply chain disruption example where a "single truth without single source" data architectures come in handy:
A tier 5 type supplier (small specialized manufacturer) goes down and you first hear about it in the trade news.
A tier 5 type supplier (small specialized manufacturer) goes down and you first hear about it in the trade news.
At this point, you, the OEM ("tier 0"), a big brand name, volume manufacturer of a high tech product, have no idea if this small company produces anything that could end up in any of your own products, assembled from some 10,000 components.
It just could be.
It just could be.
Tier 5's products end up as components is a few tier 4 suppliers' products, which should appear in tier 4's bill of materials, after which tier 5's part numbers disappear.
Tier 5 is under no obligation to give you production volumes, stock, or customer list. Neither is tier 4.
Tier 5 is under no obligation to give you production volumes, stock, or customer list. Neither is tier 4.
0⃣Little memory jog. Arriving at UofI in 1995, my first presentation (on a topic of my choosing) was about what I called "the wildcat approach to product development".
The talk was shaped by my experiences in R&D where betting on the wrong horse had become concerningly common.
The talk was shaped by my experiences in R&D where betting on the wrong horse had become concerningly common.
1⃣The idea was that product development should be split into two very separate phases with very different skills sets, but that corporate research and development no longer captured these phases, largely bc they were disintegrated activities within firms.
2⃣I called the two phases "search" and "hunt". Search is a comparatively leisurely activity, the sit idly on knoll and scan the environment phase.
It is about identifying the right target, and about coming to a mutual understanding as to why it is the right target.
It is about identifying the right target, and about coming to a mutual understanding as to why it is the right target.
Re-evaluating the battle of the Olivers based on Holmström's claim that Oliver Williamson wrote about firm as escape from the limits of the market, while Oliver Hart wrote about the market as escape from the limits of the firm. There's another Nobel paper in there somewhere.
First of all, I don't quite get the beef between two camps. They are simply writing about two different exploitation scenarios, both using the tools they perfected. Both deal with asymmetries in a Prisoners Dilemma. Williamson is longitudinal, Hart is pivotal.
Williamson is good old Carnegie engineering tradition here, while Hart adheres to the MIT modeling standard. This is all fine by itself, models are good to unearth certain facets of the world. The "my model works, so your model can't possibly unearth additional truths" is, uhh...
Decentralization is about directing information flow and replicating decision points. One of its crucial design concepts is connectivity. A central (sic) theorem on connectivity is Menger's theorem, introduced by Karl Menger, the son of Austrian economist Carl Menger, in 1927.
Decentralization differs from pure replication (same decision taken by multiple agents) and distribution (chunks of the overall decision problem assigned to multiple agents) in that it is designed into the information flow and can incorporate multiple entry/exit points.
If that's too abstract, think of a rail or road network.
