This is a brief summary of the global variables and functions in
FF. Refer to the comments in the code for more details.



Overview
--------

FF internally represents a grounded version of the planning
problem. This implies that all possible (reachable) ground facts
and ground actions are precomputed. Indeed, there is a unique
index associated to any fact and any action of a planning task.

In addition, FF converts the planning problem to a STRIPS
representation. For each conditional operator in the planning
domain, FF associates a non-empty set of STRIPS actions. In the
FF implementation operators are referred to as "op" and a strips
action as an "ef". The name for the latter is rather confusing,
as it suggests "effect". So beware, an "ef" is always a strips
*action*.

In a planning task that is already a STRIPS task, there is only
one "ef" associated to every "op".


Relevant data structures/variables
==================================

Operators (ops)
---------------

Description: operators have a number of possible strips actions
  (efs).

Data structure/variables:
- data structure for operators, defined by the struct _OpConn (ff.h:869).
- stored in global variable gop_conn[] 
- there number of operators is stored in the global variable gnum_op_conn


STRIPS actions (efs)
--------------------

Descriptions: Strips actions for each operator.

Data structure/variables:
  - data structure defined in ff.h:893
  - stored in global variabel geff_conn[] (size= gnum_ef_conn)

Elements:

- contains the following arrays with facts
  * preconditions (PC) (size= num_PC)
  * added facts   (A)  (size= num_A)
  * deleted facts (D)  (size= num_D)
  * implied efs (I) (size=num_I)

  The implied efs is only relevant in some planning tasks that
  are not formulated originally in STRIPS. This is NOT relevant
  for our assignment because the domains we are dealing with are
  STRIPS.

- there is additional data for relaxed plan extraction:

 * level: level at which this effect is executable
 * in_E: whether or not the ef is part of the relaxed plan
 * num_active_PCs: active preconditions (preconditios that have
   already been achieved)

- in_plan: whether or not the effect is part of the current plan
  

Facts
-----
Description: atomic facts of the planning problem

Data structure/variables:
  - data structure defined in ff.h:930
  - stored in global variabel gft_conn[] (size= gnum_ft_conn)

Elements:

- PC (size= num_PC): Array of effects that have this fact as a
  precondition
- A (size= num_A): Array of effects that add this fact.
- D (size= num_D): Array of effects that delete this fact.

Plans
-----

While performing the EHC, each time a new better state has been
found, the chunk of actions that were used to reach such a state
is stored in the gplan_ops[] global variable. This is done by the
extract_plan_fragment function (search.c:666).


Relevant Functions
==================

Functions that you will probably have to look into for the
assignment are:

print_op_name (output.c:701): prints an operator whose index is
given as argument.

do_enforced_hill_climbing (search.c:154): main EHC search function. Other
functions are called by this function. Refer to the code for
details.


get_1P_and_H (relax.c:111): computes a relaxed plan and returns heuristic value.

build_fixpoint (relax.c:339): builds the relaxed graph.