Swarm Intelligence-based Approach based on Smelling Capacity of Bloodhound to Solve Combinatorial Optimization Problem

Abstract

Background: Swarm Intelligence is one of the most fascinating fields to work with as it is easily relatable to the real world behaviour of animals. Combinatorial optimization problems are the complex computational problems to be solved, having multiple solution techniques. The bloodhound approach is the swarm intelligence-based approach which derives it's motivation from the real world smelling capability of the bloodhound breed of dogs.
Objectives: There are two main objectives of this technique. First, good solution quality in terms of distance for the traveling salesman problem (testbed problem). The second, to get the solution faster than the existing approach.
Methods: This bloodhound-based approach is made up of three other methods. In the first method, the initialisation of the data members is done. If we change the initial values of few parameters, then the final outcome may vary. The second one is the actual solution construction method which actually builds the solution. The third method updates the odour image of the bloodhound based on the solution quality found.
Results: The computational result shows the promising outcome considering solution quality. The quality - time trade- off is always there and one can always work further in that direction.
Conclusions: This approach is a swarm intelligence based approach which depends on the behaviour of the bloodhound breed of dogs. The bloodhound has got the highest smelling capability among all the other breeds. They can also remember the odour of the image for longer amount of time. The approach is tested on the traveling salesman problem and it provides promising solution quality with respect to other existing approaches, by compromising the time parameter little. One can work further in the direction to overcome quality - time trade-off.