1. Run the PSO implementation as a single swarm, optimizing the F6 function. Note that the run file example in Listing 4.36 is for two swarms, so make sure you have an appropriate run file for a...

1. Run the PSO implementation as a single swarm, optimizing the F6 function. Note that the run file example in Listing 4.36 is for two swarms, so make sure you have an appropriate run file for a single swarm. Try two population sizes. Turn in and discuss your results.

2. Run the PSO implementation using two swarms to optimize the G1 function. Try two different population sizes for each of the two swarms (four combinations total). Turn in and discuss your results.

3. In a single-layer neural network with n processing elements (PEs), how many unique weights are possible if the only restriction is that no self-feedback connections are allowed? How many are possible if it is also specified that weights are symmetric, that is, wji = wij?

4. Show that a hidden layer doesn’t change (improve) network performance if all PEs (hidden and output) have linear activation functions.