ss.industrialbenchmark.industrialbenchmark.1.1.0.source-code.sim.properties Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of industrialbenchmark Show documentation
Show all versions of industrialbenchmark Show documentation
A novel reinforcement learning benchmark, called Industrial Benchmark, is introduced.
The Industrial Benchmark aims at being be realistic in the sense, that it includes a variety
of aspects that we found to be vital in industrial applications. It is not designed to be an
approximation of any real system, but to pose the same hardness and complexity.
# Properties for IndustrialBenchmark
# ===================================
# benchmark params:
SEED=1
SIM_STEPS=1000
# weighting dynamics in reward
CRD=3
# weighting operationalcosts in reward
CRE=1
# weighting goldstone in reward
CRGS=25
# setpoint properties (assuming 5-10 seconds/ steps):
STATIONARY_SETPOINT = 100.0
MAX_CHANGE_RATE_PER_STEP_SETPOINT=1
MAX_SEQUENCE_LENGTH=100
SETPOINT_STEP_SIZE=1
# observation variable boundaries + initial values
Velocity_MAX=100
Velocity_MIN=0
Velocity_INIT=0
Gain_MAX=100
Gain_MIN=0
Gain_INIT=0
Shift_MAX=100
Shift_MIN=0
Shift_INIT=0
SetPoint_MIN=0
SetPoint_MAX=100
Dynamics_MIN = 0
Dynamics_MAX = 200
OperationalCostsConv_MIN=0
OperationalCostsConv_MAX=1000
RewardTotal_MIN = -6500
RewardTotal_MAX = 0
# dynamics parameters
DBase=30000
DGain=0.01
DVelocity=5
DSetPoint=100
# cost weighting parameter
CostSetPoint=2.0
CostGain=2.5
CostVelocity=4
ConvArray=0.11111,0.22222,0.33333,0.22222,0.11111,0,0,0,0,0
# step sizes
STEP_SIZE_VELOCITY=1.0
STEP_SIZE_GAIN=10.0