A method to optimize evacuation instructions

In this paper, an optimization method is presented for instructions to evacuate by car the population of a region threatened by a hazard. By giving the optimized instructions to the evacuees, traffic conditions and, therefore, the evacuation efficiency are optimized. The instructions, containing a departure time, a safe destination, and a route, are created using an optimization method consisting of two phases: the generation of the search space and the algorithm AES(+) evacuation, a version of ant colony optimization. The main contributions of the proposed optimization method are the unique approach to generate the search space in which network degeneration is taken into consideration, the possibility to optimize instructions under the assumption of both full and partial compliance of the evacuees with the instructions, and the flexibility in the sense that the user of the method can define his or her own objective function and choose a suitable traffic simulation model. The paper contains a comprehensive case study. The case study shows that the effectiveness of the optimized instructions is more than doubled when compared with the effectiveness of instructions set up by straightforward rules (like evacuating to the nearest destination using the shortest route). Further, the case study shows that the number of arrivals under optimized, but possibly sub-optimal instructions is equal to at least 90% of the theoretical upper bound on this number of arrivals.