A simple example of a deterministic simulation. When trying out a new model it’s a very good idea to first try to simulate - with the parameter values. The simulation of the model output should be realistic and comparable with the observations. Once the model, with parameter values, provides a good simulation, then it’s time to move on to estimation.

Using another optimizer for the parameter estimation can be really useful. In this example the nlminb() function is used. The negative log-likelihood is calculated in an R function, which is then passed to the optimizer. It’s also seen how actually a robust version of the negative log-likelihood is calculated.

An example of different ways to realize stochastic simulation.

Using the tikz package circutikz in LaTeX to generate RC-digrams.