YES

Problem:
 comp(s,id()) -> s
 cons(one(),shift()) -> id()
 cons(comp(one(),s),comp(shift(),s)) -> s
 comp(one(),cons(s,t)) -> s
 comp(shift(),cons(s,t)) -> t
 comp(abs(s),t) -> abs(comp(s,cons(one(),comp(t,shift()))))
 comp(cons(s,t),u) -> cons(comp(s,u),comp(t,u))
 comp(id(),s) -> s
 comp(comp(s,t),u) -> comp(s,comp(t,u))

Proof:
 WPO Processor:
  algebra: MSum
  weight function:
   w0 = 0
   w(abs) = w(cons) = w(shift) = w(one) = w(comp) = w(id) = 0
  status function:
   st(cons) = [1, 0]
   st(abs) = [0]
   st(shift) = st(one) = []
   st(comp) = [0, 1]
   st(id) = []
  subterm penalty function:
   sp(abs, 0) = 1
   sp(cons, 1) = sp(cons, 0) = sp(comp, 1) = sp(comp, 0) = 0
  subterm coefficient function:
   sc(abs, 0) = sc(cons, 1) = sc(cons, 0) = sc(comp, 1) = sc(comp, 0) = 1
  weight status function:
   ws(comp) = pol
   ws(abs) = ws(cons) = ws(shift) = ws(one) = ws(id) = max
  precedence:
   comp > cons > abs ~ shift ~ one ~ id
  problem:
   
  Qed