Следва точно решение на задачата на Python:

clone = lambda m: [ x[:] for x in [y[:] for y in m] ] # deepcopy



def R( m, k, d=-1):                 #rotate right

    o= clone(m)

    o[k] = [ m[k][(a+d)%len(o[k])] for a in range(len(o[k]))]

    return o

    

def L( m, k): return R(m,k,1)       #rotate left



def U( m, k, d=1):                  #rotate up

    o= clone(m)

    for a in range(len(o)):

        o[a][k]=m[(a+d)%len(o)][k]

    return o

    

def D( m, k): return U(m,k,-1)      #rotate down



def gen(m, m_goal, (F1, F2), D,queue):

    """ generate all 1 move next states from current state m using F1,F2"""

    for x in range(D):

            for F in (F1, F2):

                cmd=F.__name__ + str(x+1)

                next = F(m, x)

                if m == m_goal:     # solved!!!

                    print >> open("SHIFTING.OUT","w"), " ".join(queue[0][1])

                    return True

                queue.append( (next, queue[0][1] + [cmd] ) )



def solve(m, m_goal, M, N ):

    """solve by breadth first searching all posibilities."""

    queue=[ (m, []) ]; 

    while True:

        if gen(m, m_goal, (U,D), M, queue):

            return

        if gen(m, m_goal, (L,R), N, queue):

            return

        del queue[0]

        m = queue[0][0]



if __name__ == "__main__":        

    lines = open("SHIFTING.INP").readlines()

    M,N,C = map(int, lines[0].split())

    m=[]; m_goal=[]

    for a in lines[1:N+1]:

        m.append( map(int, a.split()) )

    for a in lines[N+1:]:

        m_goal.append( map(int, a.split()) )

    

    solve( m, m_goal, M,N)