Probabilistic modal {mu}-calculus with independent product

The probabilistic modal {\mu}-calculus is a fixed-point logic designed for expressing properties of probabilistic labeled transition systems (PLTS’s). Two equivalent semantics have been studied for this logic, both assigning to each state a value in the interval [0,1] representing the probability that the property expressed by the formula holds at the state. One semantics is denotational and the other is a game semantics, specified in terms of two-player stochastic parity games. A shortcoming of the probabilistic modal {\mu}-calculus is the lack of expressiveness required to encode other important temporal logics for PLTS’s such as Probabilistic Computation Tree Logic (PCTL). To address this limitation we extend the logic with a new pair of operators: independent product and coproduct. The resulting logic, called probabilistic modal {\mu}-calculus with independent product, can encode many properties of interest and subsumes the qualitative fragment of PCTL. The main contribution of this paper is the definition of an appropriate game semantics for this extended probabilistic {\mu}-calculus. This relies on the definition of a new class of games which generalize standard two-player stochastic (parity) games by allowing a play to be split into concurrent subplays, each continuing their evolution independently. Our main technical result is the equivalence of the two semantics. The proof is carried out in ZFC set theory extended with Martin’s Axiom at an uncountable cardinal.

💡 Research Summary

The paper addresses a well‑known limitation of the probabilistic modal μ‑calculus (pμ‑calculus), namely its inability to express many properties that are naturally captured by logics such as Probabilistic Computation Tree Logic (PCTL). To overcome this, the authors extend the syntax of pμ‑calculus with two new operators: independent product (⊗) and its dual coproduct (⊕). Semantically, the independent product of two formulas φ and ψ evaluates to the product of their probabilities (a·b) at a given state, while the coproduct evaluates to 1 − (1 − a)(1 − b). These operators enable the logic to describe events that occur independently and simultaneously, a feature missing from the original calculus.

The main technical contribution is a new game semantics that matches the extended syntax. Traditional two‑player stochastic parity games model pμ‑calculus formulas by a single, linear play that alternates between the two players and a probabilistic node. The independent product, however, requires a play to split into several concurrent sub‑plays, each evolving independently according to the underlying PLTS. The authors therefore define a class of split stochastic parity games. At a split node (triggered by ⊗ or ⊕), the current play branches into two (or more) sub‑plays; each sub‑play proceeds with its own parity condition and probabilistic transitions. The overall outcome is obtained by combining the results of the sub‑plays using the same algebraic operation that appears in the formula (product for ⊗, coproduct for ⊕). Player 0 (the verifier) aims to maximise the combined probability, while Player 1 (the refuter) tries to minimise it.

To prove that the denotational semantics (the fixed‑point interpretation of formulas) coincides with the value of the split game, the authors develop a sophisticated argument in set theory. The proof proceeds by structural induction on formulas for the finite‑branching case, showing that each connective preserves the equality of the two semantics. The challenging part concerns the fixed‑point operators μ and ν, which generate potentially infinite unfolding of the game tree. Because the split operation can create infinitely many concurrent branches, the authors need a robust measure‑theoretic foundation to talk about the probability of a set of infinite plays. They adopt ZFC set theory augmented with Martin’s Axiom (MA) at the uncountable cardinal ℵ₁. MA guarantees the existence of filters on ℵ₁‑complete partial orders, which in turn yields a σ‑additive probability measure on the space of infinite branching plays. With this measure in hand, the authors can define the expected value of a split game as a limit of finite approximations and show that it satisfies the same fixed‑point equations as the denotational semantics.

The equivalence theorem—for every state s and every formula φ, the denotational value ⟦φ⟧(s) equals the value of the split stochastic parity game starting from (s, φ)—is the central result. As a corollary, the extended pμ‑calculus subsumes the qualitative fragment of PCTL: properties such as “there exists a path with positive probability” or “with probability 1 all executions eventually satisfy ψ” can be encoded using the new product and coproduct operators together with the existing modal operators.

Beyond the theoretical contribution, the paper outlines several practical implications. The split game model provides a natural operational picture for reasoning about concurrent probabilistic behaviours, which is useful for model‑checking algorithms that need to handle independence explicitly. Moreover, the authors discuss how their framework could be adapted to incorporate other probabilistic constructs (e.g., conditional probabilities) and how the reliance on MA might be weakened in future work, perhaps by restricting to countable PLTSs or by employing alternative measure‑theoretic tools.

In summary, the paper delivers a comprehensive extension of the probabilistic modal μ‑calculus, introduces a novel class of stochastic parity games that support independent concurrent sub‑plays, and rigorously proves the equivalence of denotational and game semantics under standard set‑theoretic assumptions. This work significantly broadens the expressive power of pμ‑calculus and opens new avenues for the verification of complex probabilistic systems.