YES

Problem:
 +(a(),x) -> a()
 +(b(),g(a())) -> a()
 +(0(),x) -> x
 +(x,+(y,z)) -> +(+(x,y),z)
 +(+(x,y),z) -> +(x,+(y,z))
 +(x,y) -> +(y,x)

Proof:
 Church Rosser Transformation Processor (ac):
  f5_AC(a(),x) -> a()
  f5_AC(b(),g(a())) -> a()
  f5_AC(0(),x) -> x
  AC critical peaks: joinable
  AC-KBO Processor:
   precedence:
    g > 0 > f5_AC > b > a
    weight function:
     w0 = 1
     w(0) = w(g) = 4
     w(a) = 2
     w(b) = 1
     w(f5_AC) = 0
    problem:
     
    Qed