How the day-ahead auction clears

At 12:00 CET on D-1, the day-ahead gate closes. Forty minutes later, prices are published for every hour of tomorrow in every coupled zone in Europe. In between, an algorithm called EUPHEMIA runs the auction.

EUPHEMIA stands for Pan-European Hybrid Electricity Market Integration Algorithm. The name is awful. The job it does is clean.

What the algorithm is trying to do

flowchart TB
    A([Goal<br/>maximise economic welfare<br/>across every coupled zone<br/>for every hour of tomorrow])

    B([Inputs<br/>all bids and offers<br/>cross-border capacity limits<br/>block order constraints])

    C([Output<br/>one price per zone per hour<br/>cleared volumes for each bid<br/>cross-border flows])

    B --> A --> C

    style A fill:#fed7aa,stroke:#c2410c,color:#7c2d12
    style B fill:#fed7aa,stroke:#c2410c,color:#7c2d12
    style C fill:#fed7aa,stroke:#c2410c,color:#7c2d12

Maximise economic welfare sounds abstract. In practice it means: clear the cheapest supply with the highest-value demand, hour by hour, across all coupled zones, while respecting the wires between zones.

A clean way to picture it

Forget the algorithm for a second. Picture a 24-hour spreadsheet across many zones.

For each hour and each zone, you have:

• A supply curve (all the offers stacked from cheap to expensive).
• A demand curve (all the bids stacked from high-value to low).
• A wire capacity to each neighbouring zone.

The auction asks one question: what set of prices and flows leaves the most value on the table for everyone, taken together?

flowchart TB
    Z1([Zone A<br/>cheap supply<br/>not much demand])
    Z2([Zone B<br/>expensive supply<br/>lots of demand])
    W([Wire between A and B<br/>has a capacity limit])

    Z1 -->|export power<br/>up to wire limit| Z2

    style Z1 fill:#dcfce7,stroke:#15803d,color:#14532d
    style Z2 fill:#dbeafe,stroke:#1e40af,color:#1e3a8a
    style W fill:#fef3c7,stroke:#a16207,color:#713f12

If the wire is not full, the auction equalises the price in A and B. Same number. If the wire is full, the auction gives up on equalising. Zone A stays cheap. Zone B stays expensive. The wire ships whatever it can.

That is what bidding zones split means. The wire was the limit.

Two outputs nobody talks about

EUPHEMIA also delivers two outputs that newcomers do not think to ask for, but that matter operationally.

Cross-border flows for every hour. The algorithm decides not just prices but also which zone exports how much to which neighbour, in every hour. This determines what physical flow Svenska kraftnät has to handle.

Congestion rent. When two coupled zones clear at different prices, the wire between them generates congestion rent: the cheap zone collects the cheap price, the expensive zone pays the expensive price, and the difference goes to the TSOs on each side. This is a real source of revenue for Svenska kraftnät and is meant to be reinvested in new lines.

Why it has to be one big algorithm

Every coupled zone in Europe is solved at the same time, in one optimisation. Not zone by zone. Not iteratively. One big problem.

This is because the zones are linked: a flow from Norway to Germany also affects the price in Sweden, the price in Denmark, the price in the Netherlands. You cannot solve them independently and get a consistent answer.

EUPHEMIA is the algorithm that handles this all in one shot. It runs in 12 to 17 minutes, every day, for a problem that spans 25+ countries. That is the engineering achievement hiding behind the awful name.

What can go wrong

A few real-world cases.

EUPHEMIA cannot find a solution. Very rare, but it has happened. When complex bid types and tight wire limits create an infeasible problem, the algorithm fails over to a simpler version, or to manual operator intervention.

Late delivery. If the algorithm runs long, prices come out at 12:55 instead of 12:42. Traders complain, but the markets still settle.

Decoupling. When a particular zone is temporarily disconnected (technical issue, regulatory intervention), the auction runs without it. Sweden has been decoupled briefly in the past during major events.

Next

Real bids are not just I will sell at X. There are several bid types, and they shape who wins. See Bid types: hourly, block, complex.