SciML / DifferentialEquations.jl

Multi-language suite for high-performance solvers of differential equations and scientific machine learning (SciML) components. Ordinary differential equations (ODEs), stochastic differential equations (SDEs), delay differential equations (DDEs), differential-algebraic equations (DAEs), and more in Julia.
https://docs.sciml.ai/DiffEqDocs/stable/
Other
2.87k stars 230 forks source link

Morris Global Sensitivity Analysis #95

Closed ChrisRackauckas closed 7 years ago

ChrisRackauckas commented 8 years ago

Implement a sensitivity(::ODEProblem,tspan) which applies Morris global sensitivity analysis on a ParameterizedFunction.

http://www.abe.ufl.edu/Faculty/jjones/ABE_5646/2010/Morris.1991%20SA%20paper.pdf http://www.math.ualberta.ca/~hwang/sensitivity http://discovery.ucl.ac.uk/19896/1/19896.pdf

ChrisRackauckas commented 7 years ago

Migrated to DiffEqSensitivity.jl

https://github.com/JuliaDiffEq/DiffEqSensitivity.jl/issues/1

vijaya65 commented 4 years ago

I have a question regarding the Morris method in the differential equation package. The documentation does not describe how the number of trajectories is calculated for the Morris method. In the example given for the LotkaVolterra model with 4 parameters, an unusually large number of trajectories is included in the input

m = gsa(f1,Morris(total _num_trajectory=1000,num_trajectory=150),[[1,5],[1,5],[1,5],[1,5]])

From what I have read about the method, for 4 parameters the number of trajectories would be 5x(r=8) or 40 trajectories at most. Do you know how the numbers 1000 and 150 are calculated for the Lotka Volterra model and what they represent?