In cloud computing, an automated SLA is an electronic contract used to record the rights and obligations of service providers and customers for their services. SLA negotiation can be a time-consuming process, mainly due to the unpredictable rounds of negotiation and the complicated possible dependencies among SLAs. The operation of negotiating SLAs can be facilitated when SLAs are translated into Reduced Ordered Binary Decision Diagrams (ROBDDs). Nevertheless, an ROBDD may not be optimally structured upon production. In this paper, we show how to reduce the number of 1-paths and nodes of ROBDDs that model SLAs, using ROBDD optimization algorithms. In addition, we demonstrate the reduction of 1-paths via the application of Term Rewriting Systems with mutually exclusive features. Using the latter, ROBDDs can be generated accurately without redundant 1-paths. We apply the principles onto the negotiation of IaaS SLAs via simulation, and show that negotiation is accelerated by assessing fewer SLA proposals (1-paths), while memory consumption is also reduced.