Unlocking Evolutionary Game Theory: From Dynamic Equilibria to Real‑World Applications

Evolutionary Game Theory

Traditional game theory assumes (1) complete rationality and (2) complete information.

Unlike traditional game theory, evolutionary game theory does not require participants to be fully rational or to have complete information. It combines game‑theoretic analysis with dynamic evolutionary processes, focusing on dynamic equilibria rather than static ones. The theory originates from biological evolution.

Why Introduce Evolutionary Thinking into Game Theory?

(1) Influence on biology. In biology, strategies correspond to genes and payoffs to fitness; the key difference from economics is non‑fully rational choices .

(2) Impact on social sciences. In market competition, we need not rationally identify the optimal strategy; firms that survive are those with the strongest adaptability.

In evolutionary game theory, Evolutionarily Stable Strategy (ESS) and Replication Dynamics are core concepts. ESS describes a strategy that, once adopted by a population, cannot be invaded by alternative strategies due to limited rationality and learning. Replication dynamics describe how the frequency of a strategy changes over time, often expressed by a differential equation.

Replication dynamics can be represented by a differential equation where the variables denote the proportion of a pure strategy, its fitness, and the average fitness.

When time tends to infinity, the stability of strategies is examined. A stable state must be robust to small perturbations to qualify as an ESS. Mathematically, if a perturbation lowers x below the equilibrium, then dx/dt > 0 ; if it raises x above the equilibrium, then dx/dt < 0 , ensuring the derivative is negative at the equilibrium.

Symmetric Games in Replication Dynamics

(1) Signing Agreement Game

Assume the proportion of “Y” in the population is x and the proportion of “N” is 1‑x. The replication dynamics lead to a stable state where x = 1, and this state is an ESS, meaning all participants eventually choose “Y”.

(2) General Two‑Player Symmetric Game

(3) Coordination Game

The replication dynamics equation for the coordination game can be expressed as a system of differential equations (details omitted for brevity).

(4) Hawk‑Dove Game

Let x be the proportion adopting the “Hawk” strategy and 1‑x the “Dove” strategy. The replication dynamics equation is dx/dt = x(1‑x)(E_H‑E_D) , where E_H and E_D are the expected payoffs. Depending on the relative values of conflict loss e and resource value v, the system reaches different stable states. When e is much larger than v, the ESS is a mixed strategy; when e = v, both pure strategies can be stable; when e < v, the “Hawk” strategy dominates.

(5) Frog‑Calling Game

Let m and P be the probability of successful mating, and z the opportunity cost (energy, risk, etc.). The replication dynamics equation is dx/dt = x(1‑x)[(m‑z)‑P(1‑x)] . When (m‑z) > 0, a proportion (m‑z)/(1‑P) of males will call; if the benefit of “cheating” (not calling) exceeds the benefit of calling, silent strategies may dominate.

Follow the “Model Perspective” public account for more on mathematics and modeling!