Improving and Extending SMR Technique for Solving Monotonic Problems

State matrix representation (SMR) model is an efficient technique to find an optimum plan for a monotonic problem introduced in 1994 in. One of the prominent aspects of this technique is that robot action planning could be performed efficiently with simple matrix manipulation. Another important characteristic of SMR is that unlike most planning methods the actions involved in the planning process are not defined in advance but are generated at planning time. The syntax used for the actions while directly reflecting the semantics of actions, they can be easily manipulated by the planner. This paper presents extension and improvement of the SMR technique to generate more efficient plans. In this paper shortcomings of SMR technique are highlighted and an extension is provided to overcome the shortcomings. In addition, application of the improved version which we called ISMR to a new variation of the block world problem with different restriction is presented and the results are discussed.