Strathprints logo
Strathprints Home | Open Access | Browse | Search | User area | Copyright | Help | Library Home | SUPrimo

A multi-directional modified physarum solver for discrete decision making

Masi, Luca and Vasile, Massimiliano (2012) A multi-directional modified physarum solver for discrete decision making. In: Bioinspired Optimization Methods and their Applications, BIOMA 2012, 2012-05-24 - 2012-05-25, Bohinj.

[img] PDF - Draft Version
Download (140Kb)

    Abstract

    In this paper, a bio-inspired algorithm able to incrementally grow decision graphs in multiple directions is presented. The heuristic draws inspiration from the behaviour of the slime mould Physarum Polycephalum. In its main vegetative state, the plasmodium, this large single-celled amoeboid organism extends and optimizes a net of veins looking for food. The algorithm is here used to solve classical problems in operations research (symmetric Traveling Salesman and Vehicle Routing Problems). Simulations on selected test cases demonstrate that a multidirectional modied Physarum solver performs better than a unidirectional one. The ability to evaluate decisions from multiple directions enhances the performance of the solver in the construction and selection of optimal decision sequences.

    Item type: Conference or Workshop Item (Paper)
    ID code: 40480
    Keywords: discrete optimisation, physarum solver, multidirectional, Mechanical engineering and machinery, Motor vehicles. Aeronautics. Astronautics, Mechanical Engineering, Aerospace Engineering, Computational Mechanics
    Subjects: Technology > Mechanical engineering and machinery
    Technology > Motor vehicles. Aeronautics. Astronautics
    Department: Faculty of Engineering > Mechanical and Aerospace Engineering
    Technology and Innovation Centre > Advanced Engineering and Manufacturing
    Related URLs:
    Depositing user: Pure Administrator
    Date Deposited: 18 Jul 2012 16:23
    Last modified: 06 Sep 2014 20:59
    URI: http://strathprints.strath.ac.uk/id/eprint/40480

    Actions (login required)

    View Item

    Fulltext Downloads: