Such problems can also be formulated in a concise way with Clingo ASP (answer set programming):

• We define feasible combinations of depart and return trips with the resulting costs

• We minimize the costs

% ---------- INPUT ----------
d(1,10). d(2,7).  d(3,13). d(4,12). d(5,4).
r(1,5).  r(2,12). r(3,7).  r(4,10). r(5,12).

% ---------- CHOICE RULE ----------
1 { trip(I,J,D+R) : d(I,D), r(J,R), J>=I } 1.

% ---------- OPTIMIZATION ----------
#minimize { C : trip(I,J,C) }.

% ---------- OUTPUT ----------
#show trip/3.

Output: (run the code online)

clingo version 5.7.2 (6bd7584d)
Reading from stdin
Solving...
Answer: 1
trip(1,1,15)
Optimization: 15
Answer: 2
trip(2,3,14)
Optimization: 14
OPTIMUM FOUND

Models       : 2
  Optimum    : yes
Optimization : 14
Calls        : 1
Time         : 0.037s (Solving: 0.00s 1st Model: 0.00s Unsat: 0.00s)
CPU Time     : 0.000s