The Antnet Routing Algorithm A Modified Version

1
The Antnet Routing Algorithm - A Modified Version

• Firat Tekiner, Z. Ghassemlooy
• Optical Communications Research Group, The
  University of Northumbria, Newcastle upon Tyne
• S. Alkhayatt
• School of Computing Science, Sheffield Hallam
  University
• CSNDSP 20-22 July 2004

2
Contents

• Background Information
• Ant Colony Optimisation
• Agent Based Routing Algorithms
• Antnet routing algorithm
• Improvements proposed
• Simulation Environment and Results
• Concluding Remarks

3
Aims Objectives

• Designing a routing algorithm
• Scalable
• Distributed
• Intellegent
• Self - Organising
• Fault Tolerant
• Generic Network and Machine Independent

4
Routing

• In internetworking, the process of moving a
  packet of data from source to destination.
• A routing algorithm is necessary to find the
  optimal path (or the shortest path) from source
  to destination.
• Problems
• Existing algorithms are mostly Table-Based (high
  cost)
• Congestion and contention (requires traffic
  distribution)
• Requires human intelligence
• The routing algorithms that are in use are all
  static algorithms

5
Classification

• Q-Learning
• Q-routing (Boyan et al, 94) (Tekiner et al., 04)
• Dual reinforcement Q-routing (Kumar et al., 97
  01)
• Ant (software agent) based Routing Algorithms
• ABC routing (Schoonderwoerd et al., 96)
• Regular and Uniform ant routing (Subramanian et
  al., 97)
• Antnet (Dorigo et al., 98)
• Antnet (Dorigo et al., 02)
• Improved Antnet (Boyan et al., 02)
• Antnet with evaporation (Tekiner et al. 2, 04)
• Agent Distance Vector Routing (ADVR) (Amin et
  al., 01 02)

6
Comparison of Algorithms

• Antnet uses probabilistic routing tables whereas
  in Link State and Distance Vector routing table
  entries are deterministic
• Ants use less resources on the nodes
• Ants are dynamic and self organising whereas
  Distance Vector and Link State algorithms require
  human supervision
• Q-Routing does not guarantee on finding the
  shortest path always. Moreover, they can only
  find a single path, they cannot explore multiple
  paths
• In antnet stagnation is the main problem (routing
  table freezes due to selecting same path)

7
Ants In Nature - unsophisticated and simple

• Builds and protects their nests
• Sorts brood and food items
• Explore particular areas for food, and
  preferentially exploits the richest available
  food source
• Cooperates in carrying large items
• Migrates as colonies
• Leaves pheromones on their way back
• Stores information in the nature (uses world as a
  memory)
• Make decision in a stochastic way
• Always finds the shortest paths to their nests or
  food source
• Are blind, can not foresee future, and has very
  limited memory

8
Ants How do they Find Their Way?

• Ants dont know where to go initially, and choose
  paths randomly
• Ants taking the shorter path will reach the
  destinations before the those taking a long
  route. The path is marked with pheromone.
• There after the number of ants using the shorter
  path will keep increasing, since more pheromone
  is laid on the path.

9
Antnet in Detail
Positive reinforcement
Negative reinforcement
10
Three Improvements

• A. Deleting aged packets
• if PACKET AGE gt 2 x NO_OF_NODES
• then DROP PACKET
• B. Limiting the effect of r
• if (NO_OF_NODES lt 5)
• 0.1 lt r lt (1 0.1 NO_OF_NODES)
• else / if (NO_OF_NODES gt 5) /
• 0.05 lt r lt (1 0.05 NO_OF_NODES)
• C. Limiting the number of Ants in the system

11
Simulation Network
12
Simulation Parameters

• Poisson traffic distribution, with three
  different system loads low, medium and high
• 5000 packets created per node
• Average of 8 simulation runs is used for accuracy
• No packet loss due to node/link failures
• All experiments are implemented for varying ant
  creation rates, since it has a significant effect
  on the performance of the algorithm

13
Results 1
Ant rate vs. avg. delay
14
Results 2
Ant rate vs. the throughput
15
Concluding Remarks

• Detecting and removing aged packets improved
  networks performance
• Boundaries introduce reduces the effect of the
  traffic fluctuations on the solution
• No mathematical formula only constant variables
  are used
• There is a need for a second heruistic to
  optimise antnets parameters
• Stagnation is a major problem but solution does
  exists

16
Current and Future Work

• Current Work
• Stagnation problem is currently being
  investigated in different traffic models and
  network configurations.
• Evaporation 7 improvement in the performance
  of the algorithm Tekiner et al. 2, SoftCOM04
• Multiple Ant Colonies
• Aging, and Noise
• Future Work
• Hybrid Algorithm Distributed GA could be
  embedded in the proposed model Tekiner et al.,
  seminar 2
• Together with hybrid GA all constant variables
  used needs to be dynamic (currently static
  variables used).

17
Acknowledgement

• Thanks to my sponsor
• Northumbria University
• Any Questions?