The Generalized Travelling Deliveryman Problem
Zobecněný problém obchodního doručovatele
Authors
Supervisors
Reviewers
Editors
Other contributors
Journal Title
Journal ISSN
Volume Title
Publisher
České vysoké učení technické v Praze
Czech Technical University in Prague
Czech Technical University in Prague
Date
Abstract
Mobile Robot Search Problem (MRSP) je plánovací úloha, při které robot hledá předmět v předem známém prostředí. Poloha objektu není známa. Cílem úlohy je naplánovat trajektorii, která minimalizuje očekávaný čas nalezení předmětu. Na diskretizovanou variantu tohoto problému lze pohlížet jako na úlohu z teorie grafů. Tato práce přistupuje k tomuto problému tak, že reprezentuje MRSP jako nový problém nazvaný Generalized Graph Search Pro- blem with order-dependent weight, který kombinuje známé problémy Graph Search Problem (GSP) a Generalized Traveling Salesman Problem (GTSP). Nový problém lze řešit úpravou algoritmu GLNS, navrženého pro řešení GTSP, pomocí metaheuristických konceptů Adaptive Large Neighborhood Search a Simulated Annealing. S navrženými úpravami algoritmu jsme schopni efektivně řešit Generalized GSP with order-dependent weight, jak dokládáme pomocí experimentů.
The Mobile Robot Search Problem (MRSP) is a task of searching for an object in a priory known environment. The location of the object is not known. The goal is to find a trajectory that minimizes the expected time to find the object. The discretized variant of the problem can be viewed as a graph theory problem. This thesis approaches the task by representing the MRSP as a novel combinatorial optimization problem named the Generalized Graph Search Problem (GSP) with order-dependent weights, which combines the Generalized Traveling Salesman Problem (GTSP) and the GSP. The novel problem can be tackled by modifying GLNS, a solver designed for solving the GTSP using metaheuristic concepts Adaptive Large Neighborhood Search and Simulated Annealing. With proposed modifications, the solver is able to efficiently solve the Generalized GSP with order-dependent weights, as we demonstrate in the experimental evaluation.
The Mobile Robot Search Problem (MRSP) is a task of searching for an object in a priory known environment. The location of the object is not known. The goal is to find a trajectory that minimizes the expected time to find the object. The discretized variant of the problem can be viewed as a graph theory problem. This thesis approaches the task by representing the MRSP as a novel combinatorial optimization problem named the Generalized Graph Search Problem (GSP) with order-dependent weights, which combines the Generalized Traveling Salesman Problem (GTSP) and the GSP. The novel problem can be tackled by modifying GLNS, a solver designed for solving the GTSP using metaheuristic concepts Adaptive Large Neighborhood Search and Simulated Annealing. With proposed modifications, the solver is able to efficiently solve the Generalized GSP with order-dependent weights, as we demonstrate in the experimental evaluation.