operating. Therefore, the solution option set to the local solver. After 70 hours of operation, the local optimal
solution of the model is obtained by Lingo, and the results are as follows:
The distribution routes of the 3 tankers are respectively: 0-1-4-6-9-0; 0-8-7-5-0; 0-1-2-3-10-0.
The arrival time of each tanker to the gas stations is shown in Table 5.
From Table 5, we can learn: the working time of tanker1 is 2.42 hours, that of tanker 2 is 2.42 hours and that
of tanker 3 is 2.42 hours. The maximum working time is 2.42 hours, which means the optimal solution obtained
by Lingo is 2.42 hours.
The amount of refined oil distributed for each gas station is shown in Table 6.
Next, according to the heuristic algorithm coded by Matlab software, we obtain an approximate optimal distribution
route for 3 tankers: 0-1-3-2-8-0; 0-1-4-6-9-0; 0-7-5-10-0.
The arrival time of each tanker to the gas station is shown in Table 7.
From Table 7, we can learn that the working time of tanker 1 is 2.21 hours, which of tanker 2 is 2.42 hours
and that of tanker 3 is 2.37 hours. The maximum working time is 2.42 hours, which means the optimal solution
obtained by heuristic algorithm is 2.42 hours.
The amount of refined oil per tanker distributed for each gas station is shown in Table 8.
By comparison, we find that the approximate optimal solution obtained by the heuristic algorithm is approximately
the same as the local optimal solution obtained by Lingo. But the local optimal solution of Lingo is better
in terms of eliminating the working time difference of each tanker. But it needs a long time to obtain local
optimal solution by Lingo, which cannot satisfy the requirement to obtain the optimal solution in short time. The
operation time is greatly shorter by using heuristic algorithm coded by Matlab than using Lingo software to
solve the integer programming model.