Quincus

Journal

Notes on decision-making in physical networks.

Essays from the Quincus team on stochastic optimization, pricing, network design, and the engineering of systems that run live networks.

Journal

New essays by email.

One email when a new essay is published. Nothing else.

By subscribing you agree to our privacy policy.

Sovereign networks

Sovereign supply networks: why nations are becoming operators

States spent decades treating logistics as a private-sector detail. That era is ending, and the new one demands operator-grade capability at national scale.

Jul 7, 20265 min read
Sovereign networks

Food security corridors and the logistics of resilience

Food security is usually framed as production and procurement. The binding constraint, especially under stress, is more often the network in between.

Jun 30, 20265 min read
Agriculture

Perishable networks: agriculture and the value of hours

In fresh supply chains, time is not a service metric. It is the inventory itself, depleting from the moment of harvest.

Jun 16, 20264 min read
Healthcare logistics

Healthcare logistics: service levels where failure is not an option

When the cost of a stockout is measured in patient outcomes, the entire logic of network optimization changes shape.

Jun 9, 20264 min read
Healthcare logistics

Cold chain logistics: optimizing for time, temperature, and truth

Cold chain adds a second clock to every shipment. The freight is racing its own biology, and the network has to be designed for that race.

Jun 2, 20264 min read
Freight forwarding

Freight forwarding margins: where optimization pays first

Forwarding is a margin business built on thousands of small decisions per day. That structure tells you exactly where optimization pays back fastest.

May 26, 20264 min read
Perspective

Why supply chain AI fails in production

The gap between a compelling pilot and a system running a live network for years is where most supply chain AI dies. The causes are specific and avoidable.

May 19, 20265 min read
Engineering

Agentic AI in logistics: autonomy with guardrails

Agents that act, not just recommend, are coming to physical networks. The design question is not capability but the shape of the boundary around it.

May 5, 20265 min read
Engineering

The data you need is worse than you think, and that is fine

Waiting for clean data before deploying optimization is a decision to wait forever. Well-built systems are designed for the data networks actually have.

Apr 28, 20264 min read
Engineering

Predictive ETAs: what accuracy actually requires

An ETA is a forecast wearing a timestamp costume. Treating it as the distribution it really is changes what you build and what you promise.

Apr 14, 20264 min read
Engineering

Estimation and ground truth: closing the loop

An optimizer is only as honest as the estimates feeding it. The discipline of continuously reconciling model against reality is where production systems are won.

Apr 7, 20265 min read
Network design

Hub location under demand uncertainty

Facility decisions outlive every forecast used to justify them. Robustness across demand futures, not optimality in one, is the correct objective.

Mar 31, 20264 min read
Network design

Digital twins that earn their keep

Most digital twins are expensive mirrors. The useful ones are decision engines with a rendering layer, and the difference is decided in the architecture.

Mar 17, 20265 min read
Network design

Network design is a portfolio problem

A network is a set of long-lived bets on where flow will be. Portfolio theory, not map drawing, is the right frame for making them.

Mar 3, 20265 min read
Pricing and yield

Pricing under uncertainty: why margins leak at the quote

Every quote is a bet on future costs. Price on expected cost alone and you have sold an insurance policy for free.

Feb 24, 20264 min read
Pricing and yield

Cost to serve: the number most networks cannot produce

Ask what a specific shipment cost to move and most organizations can only answer with an allocation. The real number changes pricing, sales, and strategy.

Feb 10, 20265 min read
Air cargo

Capacity is perishable: the airline economics of freight

The moment a departure leaves, its unsold capacity is worth exactly zero. That single fact should organize everything about how cargo capacity is sold.

Jan 27, 20264 min read
Air cargo

Offer management: the difference between quoting and pricing

A quote answers "what does it cost." An offer answers "which of the things we could sell this customer is best for both sides." The distinction is worth points of margin.

Jan 13, 20264 min read
Air cargo

Dynamic pricing in air cargo: from rate sheets to live yield

Air cargo still prices much of its capacity off static rate sheets while demand, capacity, and competition move hourly. The gap is measured in yield.

Jan 6, 20265 min read
Maritime

What a patent on stochastic maritime optimization covers

Quincus holds a granted US patent on stochastic maritime and supply chain optimization. Here is what that protects, in plain language.

Dec 16, 20254 min read
Maritime

Port congestion is a queueing problem wearing a logistics costume

Congestion at ports follows queueing dynamics that are well understood and routinely ignored. The nonlinearity is the whole story.

Dec 2, 20255 min read
Maritime

Feeder vessel routing: small ships, hard math

Feeder networks are where maritime economics get decided at the margin, and where the routing problem is at its least forgiving.

Nov 18, 20255 min read
Routing and sequencing

The economics of empty miles

Empty repositioning is treated as waste to be minimized. It is better understood as an option premium, and priced accordingly.

Nov 4, 20254 min read
Routing and sequencing

Routing under uncertainty: sequencing when inputs move

The routing problem taught in textbooks assumes the world holds still while the vehicle moves. Real networks route through a world in motion.

Oct 21, 20255 min read
Stochastic methods

Regime switching: when networks change their own rules

Physical networks do not have one statistical personality. They switch between calm and stressed regimes, and models that ignore the switch fail precisely when it matters.

Oct 7, 20255 min read
Stochastic methods

Monte Carlo simulation as an operating tool

Simulation is usually treated as a study you commission once. Run continuously against live state, it becomes something else: an early warning system.

Sep 23, 20255 min read
Stochastic methods

Expected value is not a plan

Averages are useful for reporting and dangerous for deciding. Scenario trees restore the structure that averaging destroys.

Sep 9, 20254 min read
Stochastic methods

The case for stochastic optimization in logistics

Stochastic optimization is not an academic refinement of deterministic planning. It is a different question, and the one operators are actually asking.

Aug 26, 20255 min read
Stochastic methods

Why deterministic plans fail in physical networks

Every plan built on single-point forecasts carries a hidden assumption: that the world will hold still. Physical networks never do.

Aug 12, 20255 min read

If your network makes decisions under uncertainty, we should talk.

We work with a small number of operators at a time. Tell us what your network is optimizing for.

Start a conversation