Fig 1.
Method.
Fig 2.
Simulation and EA integration.
Fig 3.
Traffic light components.
Fig 4.
Solution representation.
Fig 5.
Evolutionary algorithm flow chart.
Fig 6.
Cycle and offset propagation.
Table 1.
Probability of mutation operators PCtO and PGt and mutation probability Pm and per signal.
Fig 7.
Distribution of home coordinates.
Fig 8.
Distribution of main activities coordinates.
Table 2.
Population per District (Census 2010 [38]).
Fig 9.
Mobility scenarios S124h, S2M and S2A.
Fig 10.
Saturation flows.
Fig 11.
Mean travel time over generations in experiments E3,E4 and E4DVM.
Fig 12.
Mean travel time of best solutions in experiments E3,E4 and E4DVM.
Table 3.
Mean travel time, standar deviation and interquartile range of best solutions in experiment E3,E4 and E4DVM.
Fig 13.
Expected number of mutated signals with constant and varying mutation schedules (E4,E4DVM).
Fig 14.
Mean travel time over generations, scenarios S2M and S2A (E4, E4DVM).
Fig 15.
Effect of crossover operator on mean travel time.
Table 4.
SMAC solutions: EA parameters.
Table 5.
Expected number of mutated signals by SMAC configuration and E4DVM.
Fig 16.
DMQ scenario flow from 07:30AM to 08:30AM.
Fig 17.
Signals corridor S1-S15.
Fig 18.
Signals corridor S1-S3 flow from 07:30AM to 08:30AM.
Fig 19.
Cycle length of the best solutions for experiment E4DVM scenario S2M.
Annotated heatmap: solutions by row (associated travel time in the row label), variables by column (one per signal), values by color (see color legend).
Fig 20.
Offset times of the best solutions for experiment E4DVM scenario S2M.
Fig 21.
Hierarchical clustering of cycle length of the best solutions for experiment E4DVM scenario S2M.
Fig 22.
Geolocation of signal clusters, experiment E4DVM scenario S2M.
Fig 23.
Geolocation of signal clusters, experiment E4DVM scenario S124h.
Table 6.
Car distribution (fuel = gasoline and weight ≦ 2Tons).
Table 7.
Emissions of best solutions on scenario S124h.
Table 8.
Emissions of best solutions on scenario S2M.
Fig 24.
Change in CO2 emissions.
Fig 25.
Variables to detect green phase crossing.
Fig 26.
Best individual (lowest travel time) cluster No.6 (a) and cluster No. 9 (b) for south-north direction, experiment E4DVM, scenario S2M.
Table 9.
Green waves cluster 6, total agents = 253.
Table 10.
Green waves cluster 9, total agents = 1400.