
1. Arrange and allocate memory to the parameters for the new prob-
lem. In the example provided on the web site, the connection
weights are arranged into two weight matrices in accordance with
the topology of the backpropagation network.
2. Change the getWeight() function source code to correspond with
the new parameter arrangement.
3. Write a new evaluation function to replace the one (bp.c) that is
provided.
4. Modify the final fitness calculation source code according to the
meaning of the return value of the new evaluation function. In
the existing bp() function, the return value is the mean sum-
squared error, not the fitness. This can be incorporated ...