YES

TRS:

  f(f(a())) -> f(g(n__f(n__a())))
  f(X) -> n__f(X)
  a() -> n__a()
  activate(n__f(X)) -> f(activate(X))
  activate(n__a()) -> a()
  activate(X) -> X

max/plus interpretations on N:

  f_A(x1) = max{7, 4}
  f#_A(x1) = max{4, 2}
  a_A = 3
  a#_A = 1
  g_A(x1) = max{8, 2}
  g#_A(x1) = max{3, 0}
  n__f_A(x1) = max{6, 5}
  n__f#_A(x1) = max{0, 4}
  n__a_A = 1
  n__a#_A = 0
  activate_A(x1) = max{3, 1 + x1}
  activate#_A(x1) = max{3, 5}

precedence:

  activate > a > f > g = n__f = n__a