Teradata has a cost-based optimizer. It does not have a rule-based optimizer, so you don't have an option to specify that.
The main difference is that a cost-based optimizer first tries to calculate the cost of each of the various paths that it can take to solve the SQL. The cost is usually in terms of CPU time and/or number of IO's. It then decides which path is the least costly option.
A rule-based optimizer operates on a set of "rules" that dictate that it will take a certain path under certain conditions in the SQL. It may even have "hints" that you can provide to the optimizer to help it decide how to solve the query.
Teradata's cost-based optimizer does not allow for hints. One of the key components used in order for Teradata to assign the correct cost is to gather statistics on the tables that the optimizer can use to help figure out how much each potential path will cost. If you do not have these statistics, it could result in the optimizer choosing a less than optimal path.
