.
            It is very strange formulation.  It differs very much from standard Linear Programming formulations.
            I will explain it in my solution below.  I suspect that there is/(there are)  error/(errors)  in the formulation.
            But I will solve the problem  "as is",  exactly as it is given.
Let S = # of Soldiers; T = # of Tarzans, D = # of Dogs.
The objective function is the production cost  C(S,T,D) = 50*S + 30*T + 20*D dollars to minimize.
Since we want to minimize, the constraints "from the upper" do not play any role.
Concretely, the constraint on capacity does not play any role, as well as the constraint on the number of Dogs.
The only constraint which really matters is the constraint
    S + T >= 400.
But since we want to minimize the production cost, we take T = 400 (more cheaper than S) and S = 0.
We also put D = 0.
So, the solution to the given problem is  T = 400, S = 0, D = 0.
It satisfies all conditions (constraints) and provides the minimum cost  C(0,400,0) = 30*400 = 12000 dollars.    ANSWER