An integrated integer programming model with a simulated annealing heuristic for the carrier vehicle traveling salesman problem

We investigate a carrier vehicle traveling salesman problem (CVTSP) to visit n-targets in Euclidean space. Each target is visited by a smaller vehicle (e.g., UAV) consecutively, launched from and landed to a carrier. The smaller vehicle and the carrier must synchronize during both launching and landing. We present a mixed-integer second-order cone programming model with TSP-like constraints to determine the best target visiting sequence and the associated launch and land coordinates. A simulated annealing algorithm is implemented to shorten the computational time. The computational experiments indicate the simulated annealing algorithm calculates good CVTSP solutions within a reasonable amount of time.