Figure 2

Start points and gradient convergence of the optimization procedure for the gene expression example: Red pluses show the potential start points. We use 10,100, and 1000 start points in case (a), (b), and (c), respectively. The markers that are connected by lines show the iterative steps of the gradient convergence while the dashed blue line shows the true values of the parameters. We chose K=5,R=100 and assume that the parameters are in the range [10⁻⁵,10³].