A Hybrid Method to Solve the Multi-UAV Dynamic Task Assignment Problem

. 2025 Apr 16;25(8):2502. doi: 10.3390/s25082502

Algorithm 2 Initial . Offline Tasks-UAVs Assignment

Input: N: Number of UAVs.

M: Number of tasks.
Q: Number of new tasks.
$n C$ : Number of clusters.
$m p S$ : Width of the simulation map.
$d i s t$ : Task distribution, which can be random, concentrated, or predetermined.

Output:

U: UAVs after updating their assigned task list upon generating the offline assignment.

1:
Generate a set of tasks $T = {T_{1}, \dots, T_{M}}$ and set of new tasks $n T = {n T_{1}, \dots, n T_{Q}}$ based on the given distribution $d i s t$ , by calling Algorithm 4.

$T \leftarrow$ Algorithm 4 ( $M, Q, m p S, d i s t$ )
2:
Define a set of UAVs $U = {U_{1}, \dots, U_{N}}$ , by calling Algorithm 5.

$U \leftarrow$ Algorithm 5 (N)
3:
Divide the tasks into clusters, using KMean algorithm, based on their geographical locations.

$C \leftarrow$ KMean( $T, n C$ )
4:
Distribute UAVs across clusters in a balanced manner, considering each UAV’s capacity load ( $L_{i}$ ) and task density in the respective cluster, by calling Algorithm 6.

$C \leftarrow$ Algorithm 6 ( $N, M, n C, C$ )
5:
Assign tasks to UAVs by locally running the Distributed CBBA Algorithm 9 in each cluster to maximize s given in Equation (1), then update the assigned task lists ( $a_{i}$ ) for each $U_{i}$ .

$U \leftarrow$ Algorithm 9 ( $U, T$ )