Quincus

Live solver

The engineunderneath the network.

Quincus builds the optimization and simulation core that physical networks run on. Ports, air cargo, freight, and healthcare logistics. Sector platforms sit on top. The intelligence stays in one place.

Nodes 00·Edges 000·Regime NOMINAL·Gap 15.00%

Why this, why now

Workflow software is being commoditized. Optimization under uncertainty is not.

Interfaces got cheap.

Any competent team can rebuild a logistics screen in a quarter. The visible layer of enterprise software is no longer where value gets defended.

The hard part moved.

What does not rebuild quickly is an engine that holds its answer when a port closes, a lane reprices, or demand breaks its own distribution.

Production is the proof.

Optimization that has never met a live network is a research result. Ours has run continuously against real operations for 11 years.

For national and sovereign networks, the argument is different. Read the sovereign brief.

Capabilities

EVERY CAPABILITY PAGE RUNS A LIVE VERSION OF THE METHOD IN YOUR BROWSER.

One engine. Six problem classes.

Routing and sequencing.

Pickup and delivery with time windows, solved with a bound rather than a guess.

Column generation loop between restricted master problem and pricing subproblemRestricted masterproblemPricing subproblemESPPRCduals πnegative reduced cost columnsNo negative column · lower bound proven
Read the method

Pricing and yield.

Network revenue management for capacity that expires.

Conversion probability and expected revenue against price with optimal price and rate cardoptimal pricerate card1.00.0R maxpricePr(win)expected revenue
Read the method

Cost to serve.

Shipment-level cost decomposition and true marginal capacity value.

Illustrative cost to serve waterfall from revenue to contribution marginRevenue+100%Linehaul-34%Handling-12%Sortation-8%Dwell-5%Last mile-18%Customs-3%Exception recovery-6%Contribution margin+14%
Read the method

Stochastic layer.

Regime switching, jump diffusion, and tail dependence feeding distributions to the optimizer.

Three regime hidden Markov state machineNormalElevatedDisrupted
Read the method

Estimation and ground truth.

Eleven years of predicted versus realized, scored continuously.

Reliability diagram comparing calibrated and uncalibrated estimators against perfect calibration0.00.20.40.60.81.00.00.20.40.60.81.0perfect calibrationuncalibrated baselinecalibrated estimatorpredicted probabilityobserved frequency
Read the method

Network design.

Where capacity goes, decided against a scenario set rather than a forecast.

Efficient frontier of candidate network designs plotted against expected cost and service level at the tenth percentile0.000.250.500.751.000.000.250.500.751.00dominatedselected designdominatedexpected costservice level at 10th percentile
Read the method

Evidence

The numbers that are hard to copy.

0
years continuous production
0
countries operating footprint
0
modules live in production
Patented
granted US patent

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