Using Stable Model Semantics (SMODELS) in the Causal Calculator (CCALC) Semra Do˘ ganda˘ g*, F. Nur Alpaslan*, Varol Akman** * Department of Computer Engineering Middle East Technical University Ankara, TR-06531 Turkey semra, alpaslan @ceng.metu.edu.tr ** Department of Computer Engineering Bilkent University Ankara, TR-06533 Turkey akman@cs.bilkent.edu.tr ABSTRACT Action Languages are formal methods of talking about actions and their effects on fluents. One recent approach in planning is to define the domains of the planning problems using action lan- guages. The aim of this research is to find a plan for a system defined in the action language C by translating it into a causal theory and then finding an equivalent logic program. The plan- ning problem will then be reduced to finding the answer set (stable model) of this logic program. This planner will be added as an extension to the Causal Calculator (CCALC) which is a model checker for the language of the causal theories. I INTRODUCTION Plan generation plays an important role in AI research. A lot of work has been done on repre- senting actions and generating plans. While reasoning about action, some problems (such as the frame, ramification, and qualification problems) have to be taken into account. It has been shown [3] that these problems can be overcome to some extent by using causal knowledge. CCALC, which is a system written at the University of Texas, Austin, finds plans and reasons about the actions defined in causal theories. Programs written in the action description language C can be given as input to the CCALC system. CCALC translates them to the corresponding causal theory. The C language can express indirect effects, implicit preconditions, action preconditions, nondeterminism, concurrent actions, and noninertial fluents. CCALC gets the equivalent propo- sitional representation of the problem by literal completion and finds the models of the system by using the available satisfiability solvers. (A user of CCALC currently has the choice of using either “relsat” [13] or “sato” [4] as the satisfiability solver). Another perspective to planning is to use logic programming. The language of logic programming offers a reasonably expressive way, enabling us to describe the effects of actions. Also there is a