“…g,n (x, y) = max0 ≤ j ≤ n d(g i j g i j−1 • • • g i 1 (x), g i j g i j−1 • • • g i 1 (y))which is equivalent to d. For every ϕ ∈ C 0 (X), n ∈ N and ε > 0, take the averageS(X, d, G, ϕ, ε, n) = 1 m n |g| = n S(X, d g,n , S g,n ϕ, ε)where S(X, d g,n , S g,n ϕ, ε) is as defined in(5) andS g,n ϕ(x) = n j=0 ϕ(g i j g i j−1 • • • g i 1 (x)).ThenP (X, d, G, •, ε) : ϕ → lim sup n → +∞ 1 n log S(X, d, G, ϕ, ε, n)is an ε-pressure function. This motivates the following notion of upper metric mean dimension with potential for the semigroup action S. Given ϕ ∈ C 0 (X), the upper metric mean dimension with potential, of the action S of G on X and ϕ, is given by mdim M (X, d, G, ϕ) = lim sup ε → 0 + P (X, d, G, ϕ, ε) log (1/ε) .…”