The following post explains how to do a grid search for SVM. You can do the same for decision tree by just changing the parameters.
Also, this link contains a function that does grid search on a neural network, you can also take inspiration from this function to perform the same on your decision tree.
The main idea is to use ndgrid and pass the parameters you want to optimize to ndgrid and form pairs to perform the exhaustive search using ndgrid.
