diff --git a/contribution/branches/PSO-DVRP/index.h b/contribution/branches/PSO-DVRP/index.h deleted file mode 100644 index 86bcf5589..000000000 --- a/contribution/branches/PSO-DVRP/index.h +++ /dev/null @@ -1,60 +0,0 @@ -/** @mainpage Welcome to PSO-DVRP for PARADISEO - -@section Introduction - -we propose a Particle Swarm Optimization metaheuristic (PSO) for resolving the Dynamic Vehicle Routing Problem (DVRP). This problem is the dynamic and real-time version of the conventional Vehicle Routing Problem (VRP). -In the classical VRP, the planning of tours is accomplished before the working day, whereas in the dynamic version of this problem, the planning of tours is made in a dynamic or real-time way. -The purpose is to try to insert the new customer orders in the already planned tours, when the vehicles are already on routes, and in minimizing the tours cost. - -This metaheuristic is based on the Particle Swarm Optimization metaheuritic (PSO). It constitutes a great bio-inspired model of self-organization. Our approach is adaptive in both the representation of particles, and the step of their evaluation. -Furthermore, we handle the DVRP as a mixed problem. This problem is treated as an Open Vehicle Routing Problem (OVRP) at the beginning of the working day. It consists only in optimizing the segments of routes traveled by vehicles. -At the end of working day, the problem is treated as a conventional VRP. In this case, the optimization is operated on complete tours. This means that we consider the return to the depot of the whole vehicle fleet. - -As perspective, we expect to parallelize it and scale it on grid. A possible schema of parallelization is the multi-swarm, where several swarms evolve (move) in parallel on different calculation nodes. - - -@section authors AUTHORS - -
| Dolphin project-team INRIA Futurs, 2008. | -Mostepha Redouane Khouadjia |
-