The listener's policy is the softmax at rationality β of an action's expected reward under her
belief.
- policy β L R = ProbabilityTheory.Kernel.softmax fun (u : U) (a : A) => ↑(β * ∫ (w : W), R a w ∂L u)
Used by instIsFiniteKernelPolicy, instIsMarkovKernelPolicyOfNonempty, policy_real_singleton, policy_real_lt_iff, policy_single_real_lt_iff, policy_real_of_pair, integral_policy_eq_of_integral_eq_add
Used in HardingGerstenbergIcard2025, QingFranke2015, SumersEtAl2024