next up previous contents
Next: Test Evaluation Phase Up: 8.2 Optimization Results Previous: 8.2 Optimization Results

8.2.1 Computation Efficiency

Job farming has great impact on the performance of the optimization procedure. Arising from the algorithm which is typically used by optimizers we can distinguish between two phases of the whole procedure. An iteration cycle starts with a test evaluation of a set of parameters which has been identified to deliver a better target. After an improvement was achieved the optimizer evaluates the Jacobian of the model, which is necessary to be able to search for better sets of model parameters. Each of these phases requires different capabilities of the job farming system for optimal performance.

Rudi Strasser