Learning Construction in a Simulated Environment Using Temporal Goal Sequencing and Reinforcement