An introduction to cooperative game theory¶

Game theory studies situations where our outcomes depend not only on our own decisions, but also on the choices of others.

This happens all the time: in negotiations, markets, collaborative projects, shared infrastructure, political coalitions -- and even in the everyday choice between cooperating or acting alone.

In cooperative game theory, however, the focus shifts.

Instead of asking:

"What move should I make now?"

we ask a different question. Simpler. And often more practical:

What do we gain by cooperating -- and how should we split that gain?

Imagine a group of agents who can form alliances. Alone, each agent can generate some value. Together, they can often generate much more.

But the truly interesting point is not just that cooperation creates value.

It is that different groups create different amounts of value.

Some coalitions are strong. Others are weak. Certain participants are decisive in some combinations and almost irrelevant in others.

Cooperative game theory exists precisely to describe and analyze this structure of cooperation.

Here, we do not focus on step-by-step decisions or individual strategies. We assume something else: that participants can make binding agreements¹.

The central object is no longer strategy, but the value structure of coalitions -- that is, how much each possible group can generate when it cooperates.

From there, we begin to ask questions such as:

• Who is essential for cooperation to work?
• Which agreements are stable, in the sense that no group has an incentive to leave them?
• What does a "fair" split mean in practice?
• To what extent does the outcome depend on bargaining power or voting rules?

Different questions lead to different solution concepts.

Some prioritize fairness and symmetry. Others emphasize stability and resistance to deviations. Others still capture influence or power within the group.

None of them is universally "correct". Each reflects a particular way of looking at cooperation.

An example helps make this concrete.

Imagine that building a gas pipeline could benefit several interested companies. Building it is too expensive for a single company. But when several companies join forces, the project becomes viable.

Different groups of companies can make the project viable in different ways and therefore generate different construction costs (or different net benefits).

So a natural question arises:

How should we reasonably split the cost -- or the benefit -- of that cooperation?

Cooperative game theory provides the tools to describe situations like this and to compare different ways of sharing the value created when agents decide to cooperate.

The next sections are not meant to overwhelm you with formal definitions or technical proofs.

The goal here is different.

It is to build intuition.

To understand what each concept is trying to capture, why it works the way it does, and how to interpret its results in practice.

If, by the end, you can look at an allocation and think:

"This makes sense, given what we value here: fairness, stability, or power."

then you will already have grasped the essence of cooperative game theory.