O slideshow foi denunciado.
Utilizamos seu perfil e dados de atividades no LinkedIn para personalizar e exibir anúncios mais relevantes. Altere suas preferências de anúncios quando desejar.

Hybrid Evolutionary Approaches to Maximum Lifetime Routing and Energy Efficiency in Sensor Mesh Networks

Mesh network topologies are becoming increasingly popular in battery-powered wireless sensor networks, primarily because of the extension of network range. However, multihop mesh networks suffer from higher energy costs, and the routing strategy employed directly affects the lifetime of nodes with limited energy resources. Hence when planning routes there are trade-offs to be considered between individual and system-wide battery lifetimes. We present a multiobjective routing optimisation approach using hybrid evolutionary algorithms to approximate the optimal trade-off between the minimum lifetime and the average lifetime of nodes in the network. In order to accomplish this combinatorial optimisation rapidly, our approach prunes the search space using k-shortest path pruning and a graph reduction method that finds candidate routes promoting long minimum lifetimes. When arbitrarily many routes from a node to the base station are permitted, optimal routes may be found as the solution to a well-known linear program. We present an evolutionary algorithm that finds good routes when each node is allowed only a small number of paths to the base station. On a real network deployed in the Victoria & Albert Museum, London, these solutions, using only three paths per node, are able to achieve minimum lifetimes of over 99% of the optimum linear program solution’s time to first sensor battery failure.

The link for the paper: http://www.mitpressjournals.org/doi/abs/10.1162/EVCO_a_00151#.Vv6oZmErJhE

More information on our work can be found on: http://emps.exeter.ac.uk/computer-science/wsn/

  • Seja o primeiro a comentar

Hybrid Evolutionary Approaches to Maximum Lifetime Routing and Energy Efficiency in Sensor Mesh Networks

  1. 1. Hybrid Evolutionary Approaches to Maximum Lifetime Routing and Energy Efficiency in Sensor Mesh Networks Evolutionary Computation, 2015 DOI: 10.1162/EVCO a 00151 Alma Rahat Richard Everson Jonathan Fieldsend Computer Science University of Exeter United Kingdom Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 1 / 12
  2. 2. Wireless Sensors Autonomous devices Send data to a central base station Environmental or process monitoring Industrial Heritage Pharmaceuticals Health-care Battery powered Monitor locations that are difficult to access Typically left unattended for long periods of time pictu Sensor monitoring showcase environment in Mary Rose Museum, UK Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 2 / 12
  3. 3. Mesh Network and Routing Scheme Sensors and gateway Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 3 / 12
  4. 4. Mesh Network and Routing Scheme Sensors and gateway Network connectivity map Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 3 / 12
  5. 5. Mesh Network and Routing Scheme Sensors and gateway Network connectivity map Mesh Topology: sensors send data either directly (e.g. S2 = 2, G ) or indirectly (e.g. S2 = 2, 5, G ) to the gateway Alternative routes Range extension Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 3 / 12
  6. 6. Mesh Network and Routing Scheme Sensors and gateway Network connectivity map Mesh Topology: sensors send data either directly (e.g. S2 = 2, G ) or indirectly (e.g. S2 = 2, 5, G ) to the gateway Alternative routes Range extension A routing scheme for the network R = S1, S2, S3, S4, S5 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 3 / 12
  7. 7. Mesh Network and Routing Scheme Sensors and gateway Network connectivity map Mesh Topology: sensors send data either directly (e.g. S2 = 2, G ) or indirectly (e.g. S2 = 2, 5, G ) to the gateway Alternative routes Range extension A routing scheme for the network R = S1, S2, S3, S4, S5 Maximise Average lifetime Time before the first node exhausts its battery (network lifetime) Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 3 / 12
  8. 8. Node Costs Node’s cost due to a routing scheme R: C1 =T1,G + (R2,1 + T1,G) + (R3,1 + T1,G) For all transmissions. Ti,j Transmission cost at node vi Rj,i Reception cost at node vi Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 4 / 12
  9. 9. Node Costs Node’s cost due to a routing scheme R: C1 =T1,G + (R2,1 + T1,G) + (R3,1 + T1,G) For all transmissions. Ti,j Transmission cost at node vi Rj,i Reception cost at node vi T1,G Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 4 / 12
  10. 10. Node Costs Node’s cost due to a routing scheme R: C1 =T1,G + (R2,1 + T1,G) + (R3,1 + T1,G) For all transmissions. Ti,j Transmission cost at node vi Rj,i Reception cost at node vi T1,G R2,1 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 4 / 12
  11. 11. Node Costs Node’s cost due to a routing scheme R: C1 =T1,G + (R2,1 + T1,G) + (R3,1 + T1,G) For all transmissions. Ti,j Transmission cost at node vi Rj,i Reception cost at node vi T1,G R3,1 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 4 / 12
  12. 12. Node Costs Node’s cost due to a routing scheme R: C1 =T1,G + (R2,1 + T1,G) + (R3,1 + T1,G) =u1,GT1,G + u1,2R2,1 +u1,3R3,1 For all transmissions. Ti,j Transmission cost at node vi Rj,i Reception cost at node vi ui,j Edge utilisation between vi & vj for all routes u1,GT1,G u1,2R1,2 u1,3R1,3 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 4 / 12
  13. 13. Objectives Lifetime for node vi : Li (R) = Qi Ei + Ci Radio communication current Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 5 / 12
  14. 14. Objectives Lifetime for node vi : Li (R) = Qi Ei + Ci Radio communication currentQuiescent current Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 5 / 12
  15. 15. Objectives Lifetime for node vi : Li (R) = Qi Ei + Ci Radio communication currentQuiescent current Remaining battery charge Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 5 / 12
  16. 16. Objectives Lifetime for node vi : Li (R) = Qi Ei + Ci Radio communication currentQuiescent current Remaining battery charge Maximise Average lifetime: f1(R) = 1 n n i=1 Li (R) Network lifetime: f2(R) = min i∈[1,n] Li (R) Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 5 / 12
  17. 17. Search Space Size How big is the search space? Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  18. 18. Search Space Size Number of possible loopless paths for node v3: 1 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  19. 19. Search Space Size Number of possible loopless paths for node v3: 2 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  20. 20. Search Space Size Number of possible loopless paths for node v3: 3 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  21. 21. Search Space Size Number of possible loopless paths for node v3: 4 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  22. 22. Search Space Size Number of possible loopless paths for node v3: 5 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  23. 23. Search Space Size Number of possible loopless paths for node v3: 6 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  24. 24. Search Space Size Number of possible loopless paths for node v3: 7 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  25. 25. Search Space Size Number of possible loopless paths for node v3: 7 Number of possible routing schemes: n i=1 ai ai : Number of available routes from vi to vG Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  26. 26. Search Space Size Number of possible loopless paths for node v3: 7 Number of possible routing schemes: n i=1 ai ai : Number of available routes from vi to vG 4032 solutions Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  27. 27. Search Space Size Number of possible loopless paths for node v3: 7 Number of possible routing schemes: n i=1 ai ai : Number of available routes from vi to vG 243 solutions Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  28. 28. Search Space Size Number of possible loopless paths for node v3: 7 Number of possible routing schemes: n i=1 ai ai : Number of available routes from vi to vG 243 solutions Shorter paths are expected to be energy efficient Limit the number of paths available to each node by using k-shortest paths algorithm [Yen, 1972; Eppstein, 1999] Maximum search space size: kn Quicker approximation of Pareto Front Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 6 / 12
  29. 29. Max-Min Lifetime Pruning With no limits on the number of routes per node, a linear program (LP) can be derived to maximise network lifetime [Chang et al., 2004] max min vi ∈V Li subject to: Edge utilisation, uij ≥ 0 Energy usage ≤ available charge Flow conservation Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 7 / 12
  30. 30. Max-Min Lifetime Pruning Solving LP results in best network lifetime and associated edge utilisations Remove unused edges (grey) to reduce graph Apply k-SP to extract search space Ω With no limits on the number of routes per node, a linear program (LP) can be derived to maximise network lifetime [Chang et al., 2004] max min vi ∈V Li subject to: Edge utilisation, uij ≥ 0 Energy usage ≤ available charge Flow conservation Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 7 / 12
  31. 31. Multi-Objective Evolutionary Algorithm 1: A ← InitialiseArchive() Initialise elite archive randomly 2: for i ← 1 : T do 3: R1, R2 ← Select(A) Select two parent solutions 4: R ← CrossOver(R1, R2) 5: R ← Mutate(R ) 6: A ← NonDominated(A ∪ R ) Update archive 7: end for 8: return A Approximation of the Pareto set Crossover Select paths for each node from parents Mutation Replace paths randomly from k-shortest paths for some nodes Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 8 / 12
  32. 32. Hybrid Evolutionary Approach 1 Gather connectivity map, G 2 Solve LP and erase unused edges to reduce graph, G 3 Search space pruning Apply k-SP on G to generate search space Ω Apply k-SP on G to generate search space Ω Two stages of optimisation Separate optimisation: apply MOEA on Ω and Ω ; get resulting estimated Pareto set A and A Combined optimisation Use non-dominated solutions in A ∪ A as the initial archive for combined stage Apply MOEA in the combined search space Ω ∪ Ω : resulting estimated Pareto front is A Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 9 / 12
  33. 33. Real Network: The Victoria & Albert Museum Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  34. 34. Real Network: The Victoria & Albert Museum Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  35. 35. Real Network: The Victoria & Albert Museum 1st stage: optimising in Ω and Ω separately 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) ΩΩ 30 nodes + gateway k = 10; Ω and Ω are limited to 1030 solutions each. Initial population size: 100 Mutation and crossover rate: 0.1 Number of iterations: 150, 000 (1st stage) and 500, 000 (2nd stage). Run time: 2 minutes (1st stage) and 4 minutes (2nd stage). Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  36. 36. Real Network: The Victoria & Albert Museum 1st stage: optimising in Ω and Ω separately 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) ΩΩ 30 nodes + gateway k = 10; Ω and Ω are limited to 1030 solutions each. Initial population size: 100 Mutation and crossover rate: 0.1 Number of iterations: 150, 000 (1st stage) and 500, 000 (2nd stage). Run time: 2 minutes (1st stage) and 4 minutes (2nd stage). Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  37. 37. Real Network: The Victoria & Albert Museum 1st stage: optimising in Ω and Ω separately 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) ΩΩ 30 nodes + gateway k = 10; Ω and Ω are limited to 1030 solutions each. Initial population size: 100 Mutation and crossover rate: 0.1 Number of iterations: 150, 000 (1st stage) and 500, 000 (2nd stage). Run time: 2 minutes (1st stage) and 4 minutes (2nd stage). Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  38. 38. Real Network: The Victoria & Albert Museum 1st stage: optimising in Ω and Ω separately 2nd stage: optimising in Ω ∪ Ω 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) Ω ∪ Ω ΩΩ 30 nodes + gateway k = 10; Ω and Ω are limited to 1030 solutions each. Initial population size: 100 Mutation and crossover rate: 0.1 Number of iterations: 150, 000 (1st stage) and 500, 000 (2nd stage). Run time: 2 minutes (1st stage) and 4 minutes (2nd stage). Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  39. 39. Real Network: The Victoria & Albert Museum 1st stage: optimising in Ω and Ω separately 2nd stage: optimising in Ω ∪ Ω 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) Ω ∪ Ω ΩΩ 30 nodes + gateway k = 10; Ω and Ω are limited to 1030 solutions each. Initial population size: 100 Mutation and crossover rate: 0.1 Number of iterations: 150, 000 (1st stage) and 500, 000 (2nd stage). Run time: 2 minutes (1st stage) and 4 minutes (2nd stage). Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  40. 40. Real Network: The Victoria & Albert Museum 0 100000 200000 300000 400000 500000 600000 700000 800000 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 Function Evaluations Hypervolume Single-stage vs.Two-stage Ω ∪ Ω Ω ∪ Ω Ω Ω Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  41. 41. Real Network: The Victoria & Albert Museum 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 LifetimeRemaining(years) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 EdgeUtilisation 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1.92 1.93 1.94 1.95 1.96 1.97 1.98 1.99 2.00 2.01 0.7 0.8 0.9 1.0 1.1 Average lifetime: 2 years Network lifetime: 0.7 years (node v19) Avg. Lifetime Net.Lifetime Gateway Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  42. 42. Real Network: The Victoria & Albert Museum 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 LifetimeRemaining(years) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 EdgeUtilisation 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1.92 1.93 1.94 1.95 1.96 1.97 1.98 1.99 2.00 2.01 0.7 0.8 0.9 1.0 1.1 Average lifetime: 1.76 years Network lifetime: 1.29 years (node v13) Avg. Lifetime Net.Lifetime Gateway Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  43. 43. Real Network: The Victoria & Albert Museum 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 LifetimeRemaining(years) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 EdgeUtilisation 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1.92 1.93 1.94 1.95 1.96 1.97 1.98 1.99 2.00 2.01 0.7 0.8 0.9 1.0 1.1 Average lifetime: 1.94 years Network lifetime: 1.11 years (node v21) Avg. Lifetime Net.Lifetime Gateway Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 10 / 12
  44. 44. Multipath Routing Schemes Multiple routes available for each node for sending data to the base station D routes per node (D-RS): R = R1, R2, . . . , RD Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  45. 45. Multipath Routing Schemes R1 active until node 1 expires Node 1 Node 5 Charge Time Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  46. 46. Multipath Routing Schemes R1 active until node 1 expires R2 active until node 5 expires Node 1 Node 5 Charge Time Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  47. 47. Multipath Routing Schemes R1 active for time τ1 2-RS R1 active until node 1 expires R2 active until node 5 expires Node 1 Node 5 Charge Time τ1 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  48. 48. Multipath Routing Schemes R1 active for time τ1 2-RS R2 active for time τ2 R1 active until node 1 expires R2 active until node 5 expires Node 1 Node 5 Charge Time τ1 τ2 Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  49. 49. Multipath Routing Schemes R1 active for time τ1 2-RS R2 active for time τ2 R1 active until node 1 expires R2 active until node 5 expires Node 1 Node 5 Charge Time τ1 τ2 Optimal time share linear program max(τ1 + τ2) subject to: Time share, τi ≥ 0 Remaining charge ≥ 0 Linear program solved computa- tionally for each proposed routing scheme Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  50. 50. Multipath Routing Schemes Optimising in Ω and Ω separately 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) ΩΩ Hybrid evolutionary approach Evolve 1-RS solutions in Ω and Ω separately Evolve D-RS solutions in Ω and Ω separately Evolve D-RS solutions in combined search space Ω ∪ Ω Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  51. 51. Multipath Routing Schemes Optimising in Ω and Ω separately 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) ΩΩ R1 R1, R2, R3 Hybrid evolutionary approach Evolve 1-RS solutions in Ω and Ω separately Evolve D-RS solutions in Ω and Ω separately Evolve D-RS solutions in combined search space Ω ∪ Ω Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  52. 52. Multipath Routing Schemes Optimising in Ω and Ω separately Optimising in combined search space Ω ∪ Ω 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) Ω ∪ Ω ΩΩ Hybrid evolutionary approach Evolve 1-RS solutions in Ω and Ω separately Evolve D-RS solutions in Ω and Ω separately Evolve D-RS solutions in combined search space Ω ∪ Ω Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  53. 53. Multipath Routing Schemes Optimising in Ω and Ω separately Optimising in combined search space Ω ∪ Ω 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) Ω ∪ Ω ΩΩ Hybrid evolutionary approach Evolve 1-RS solutions in Ω and Ω separately Evolve D-RS solutions in Ω and Ω separately Evolve D-RS solutions in combined search space Ω ∪ Ω Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  54. 54. Multipath Routing Schemes Optimising in Ω and Ω separately Optimising in combined search space Ω ∪ Ω 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) Ω ∪ Ω ΩΩ 98.4% Hybrid evolutionary approach Evolve 1-RS solutions in Ω and Ω separately Evolve D-RS solutions in Ω and Ω separately Evolve D-RS solutions in combined search space Ω ∪ Ω Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  55. 55. Multipath Routing Schemes Optimising in Ω and Ω separately Optimising in combined search space Ω ∪ Ω 1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Average Lifetime (years) NetworkLifetime(years) Ω ∪ Ω ΩΩ 98.4% 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 LifetimeRemaining(years) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 EdgeUtilisation 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 LifetimeRemaining(years) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 EdgeUtilisation 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 65.8% 31.3% 2.9% Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 11 / 12
  56. 56. Summary Multi-objective optimisation of routing schemes to extend battery powered mesh network lifetime Novel search space pruning based on exact solution from solving a linear program for network lifetime Two-stage evolutionary approach to better approximate the trade-off between network lifetime and average lifetime Optimal time distribution between multiple routing schemes to achieve improved network lifetime About 22% overall performance gain compared to previous results 510152025 Robustness 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 NetworkLifetime(years) 1-RS 2-RS Current Work Estimate the trade-off between network lifetime and robustness (tolerance against edge failure) Rahat, Everson & Fieldsend Max. Lifetime Routing and Energy Efficiency GECCO, July 2015 12 / 12

×