A comparative study between algorithms solving Ants Nearby Treasure Search problem for multiple treasures

There is a need for fast and resource effective algorithms that can find people in need. This can be achieved using bio-inspired autonomous agents. One problem that is similar to the one of a rescue mission is Ants Nearby Treasure Search (ANTS). In this problem, a number of ants work together toward...

Full description

Bibliographic Details
Main Authors: Taskin, Kasim, Dinler, Mustafa
Format: Others
Language:English
Published: KTH, Skolan för elektroteknik och datavetenskap (EECS) 2018
Subjects:
Online Access:http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-229748
Description
Summary:There is a need for fast and resource effective algorithms that can find people in need. This can be achieved using bio-inspired autonomous agents. One problem that is similar to the one of a rescue mission is Ants Nearby Treasure Search (ANTS). In this problem, a number of ants work together towards a common goal on a grid, for example finding a treasure. To make the problem more similar to a rescue mission, it is expanded to include multiple treasures. We name this problem Ants Nearby Treasures Search (ANTSS). ANTSS is solved using Parallel Diamond Search algorithm. We propose an addition to this solution. By changing the search origin we see if our algorithm performs better, mostly in cases where treasures are grouped up. Results show how changing the origin has a significant advantage in cases were treasures are grouped up. However, there is also huge disadvantages in general test cases. === Det finns ett behov av snabba och resurseffektiva algoritmer som kan hitta personer i nöd. Detta kan uppnås med hjälp av biologiskt inspirerade autonoma agenter. Ett problem som liknar det utav ett räddningsuppdrag är Ants Nearby Treasure Search (ANTS). I detta problem arbetar myror mot ett gemensamt mål på ett rutnät, till exempel att hitta en skatt. För att göra problemet mer likt ett räddningsuppdrag utvidgas det till att omfatta flera skatter. Vi döper det nya problemet Ants Nearby Treasures Search (ANTSS). För att lösa ANTSS används Parallel Diamond Search-algoritmen. Vi föreslår en utvidgning till denna algoritm. Genom att ändra sökningens ursprung undersöks ifall vår algoritm fungerar bättre, mestadels i fall där skatter är grupperade. Resultaten visar hur förändringen av ursprunget har en stor fördel i de fall där skatter grupperas. Det finns däremot också stora nackdelar i allmänna testfall.