par Gorrini, Vittorio 
;Dorigo, Marco
Référence Lecture notes in computer science, 1063, page (345-354)
Publication Publié, 1995
;Dorigo, Marco Référence Lecture notes in computer science, 1063, page (345-354)
Publication Publié, 1995
Article révisé par les pairs
| Résumé : | The application problem we discuss in this paper is how to translate schedules for the activity of a robot whose task is to feed some machines with cylinders of raw material into loading and routing instructions. The application can be modeled as the interplay of two combinatorial optimization problems: a particular kind of knapsack and a routing problem. The knapsack problem consists of placing the largest possible number of cylinders of raw material on the robotic platform, respecting some physical constraints. To do this we apply an evolutionary strategy algorithm to individuals which code the disposition of cylinders of raw material. The coding is obtained by a set of real values which are used by an algorithm which simulates a gravitational process to produce possible solutions. After a feasible and possibly good solution is found, this is given as input to a routing optimization routine. The routine is implemented by means of a classical genetic algorithm. The main contribution of this paper is to show how evolutionary computation can be applied to solve a relatively complex real problem. Results have shown this to be a viable approach whose main merit is to be simple and effective. |
