This module implements a distributed dual-decomposition-based gradient algorithm, for a coordinated adjustment of the traffic to inject by each demand (congestion control), and the routing (backpressure based) of this traffic in the network, to maximize the network utility enforcing a fair allocation of the resources.
Ths event processor is adapted to permit observing the algorithm performances under user-defined conditions,
including asynchronous distributed executions, where signaling can be affected by losses and/or delays, and/or measurement errors.
The time evolution of different metrics can be stored in output files, for later processing.
As an example, see the fig_sec11_4_congControlAndBackpressureRouting_dualDecomp.m MATLAB file used for generating the graph/s of the case study in the
book using this algorithm.
To simulate a network with this module, use the Online_evGen_doNothing generator.
cc_update_isSynchronous: Boolean type (default: false). true if all the distributed agents involved wake up synchronousely to update its state
cc_update_averageInterUpdateTime: Default: 10.0 Average time between two updates of an agent
cc_update_maxFluctuationInterUpdateTime: Default: 5.0 Max fluctuation in time in the update interval of an agent, in absolute time values. The update intervals are sampled from a uniform distribution within the given interval
cc_simulation_maxNumberOfUpdateIntervals: Default: 1000.0 Maximum number of update intervals in average per agent
cc_control_minHd: Default: 0.1 Minimum traffic assigned to each demand
cc_control_maxHd: Default: 25.0 Maximum traffic assigned to each demand
cc_control_fairnessFactor_1: Default: 2.0 Fairness factor in utility function of congestion control for demands
simulation_randomSeed: Default: 1 Seed of the random number generator
simulation_outFileNameRoot: Default: crossLayerCongControlBackpressureRoutingDualDecomp Root of the file name to be used in the output files. If blank, no output
signaling_isSynchronous: Boolean type (default: false). true if all the distributed agents involved wake up synchronously to send the signaling messages
signaling_averageInterMessageTime: Default: 1.0 Average time between two signaling messages sent by an agent
signaling_maxFluctuationInterMessageTime: Default: 0.5 Max fluctuation in time between two signaling messages sent by an agent
signaling_averageDelay: Default: 0.0 Average time between signaling message transmission by an agent and its reception by other or others
signaling_maxFluctuationInDelay: Default: 0.0 Max fluctuation in time in the signaling delay, in absolute time values. The signaling delays are sampled from a uniform distribution within the given interval
signaling_signalingLossProbability: Default: 0.05 Probability that a signaling message transmitted is lost (not received by other or others involved agents)
routing_fixedPacketDurationAndSchedulingInterval: Default: 1.0 Fixed slot size (synchronized all links)
routing_numTrafficUnitsOfOnePacket: Default: 1.0 One packet in one slot time is this traffic
routing_statNumSchedSlotBetweenN2PRecomputing: Default: 100 Number of scheduling slots between two N2P update and storing a stat trace
routing_maxNumberOfN2PStatisticsIntervals: Default: 15 Simulation length in number of scheduling intervals
routing_gradient_gammaStep: Default: 5.0E-6 Gamma step in the gradient algorithm