Let N be the number of ads in the newspaper, M the number in the trade magazine and T the number on TV.
Cost of advertising, $C=1000N+300M+2000T to be minimised;
City exposure: 3.5N+0.5M+5T≥50;
Suburb exposure: 3N+M+3T≥60;
N,M,T>0, because all media are to be used.
To ensure none of the media use more than 45% of C:
45% of C is 450N+135M+900T.
1000N≤450N+135M+900T, 300M≤450N+135M+900T, 2000T≤450N+135M+900T, so:
-550N+135M+900T≥0, 450N-135N+900T≥0, 450N+135M-1100T≥0 are more constraints.
The 5 constraints are shown in red, while the objective function is shown in green.
In Excel, I made columns for N, M, T and C where C values use the objective function C(N,M,T) shown above in green.
I added columns for the 5 constraints and values are calculated from the functions:
3.5N+0.5M+5T-50, 3N+M+3T-60, -550N+135M+900T, 450N-135N+900T, 450N+135M-1100T.
The starting values for N, M, T are all 1 so these values produced many negative values in the constraints columns, as expected.
The next step is take ascending values of T and find minimum values of C when the constraints for N and M are satisfied.
T=1:
3.5N+0.5M≥45, 3N+M≥57
7N+M≥90. The lines 3N+M=57 and 7N=90 intersect at 4N=33, N=8.25. If we set N=8 (has to be an integer), then M≥34. When N=7, M≥41, but according to Excel this makes M cost exceed 45% of the total. When N=9, however, M≥30, and Excel confirms that all constraints are met and C=$20000. In particular, 3 constraints (suburb exposure, 45% limit for N and M) are just met (M and N use exactly 45% each of the total).