[preface] Artificial intelligence is a field of study that encompasses computational
techniques for performing tasks that require intelligence when performed by humans.
p.3 Artificial intelligence is both an art and a science. A science is a body of proved
principles that have been abstracted from nature through processes of empirical inquiry and logical deduction. On the other
hand, an art is for the most part a collection of techniques, developed pragmatically to a sophisticated level, but not necessarily
in a logical way. The field of AI is fascinating because of this complementary art and science (Yazani 1986).
p.6 AI may well be one of the most important developments of the world.
p.10 There are two kinds of heuristics, special and general. A special heuristic is one that applies to
a particular problem, whereas a general heuristic can be applied to a wide range of problems.
p.24 An expert system is an intelligent computer program that uses knowledge and inference procedures
to solve problems that are difficult enough to require significant human expertise for their solution (Bielawski
& Lewand 1991)... An expert system can also be defined as a computer application that solves complicated problems which
would otherwise require extensive human expertise (Waterman 1985).
p.93 Humans generally consider a number of alternate strategies on their way to solving problems (Luger
2001; Rich & Knight 2003). A chess player typically considers a number of possible test moves, selecting the best one,
according to criteria such as the opponent's possible responses or the degree to which various moves support a certain global
game strategy.
p.95 Human beings do not use exhaustive search: a chess player mostly examines only those
moves that he finds effective according to his own experience... Human problem solving seems to be based on judgmental
rules, which guide the search to those portions of the state space that seem somehow "promising". These rules, known
as heuristics, constitute one of the central topics of AI research. A heuristic is a strategy for selectively
searching a problem space. It guides our search along lines that have a high probability of success, avoiding wasted
efforts... Although it does not always guarantee an optimal solution to a problem, a good heuristic can, and should, come
close most of the time. Most important, it employs knowledge about the nature of the problem to find an efficient solution.
If a state space search gives us a means for formalizing the problem-solving process, heuristics allow us to infuse
that formalism with intelligence.
p.99 There are many ways of introducing knowledge to the process of formulating a problem in terms of states
and operators (Firebaugh 1988; Russell & Norvig 2002). We can use appropriate knowledge to determine the node
to be expanded in the next step. Normally, the knowledge to make this decision is provided by an evaluation function
that returns a number denoting the desirability of expanding a particular node. When the nodes are so ordered that the one
with the best result for the evaluation function is expanded first, the resulting strategy is called best-first search.
p.100 While solving a problem, we have to compromise between finding the optimum solution through
exhaustive search and finding a satisfactory solution with minimal search effort (Russell & Norvig 2002).
p.100 One has to keep the following things in mind while formulating heuristics. Prefer the most informed
heuristic evaluation function... Good heuristic evaluation functions are efficiently computed. If we have
to spend as much time for evaluating a state as it would take to expand fifty others, consider a weaker heuristic function.
p.104 Best-first search is a general algorithm for searching any state space graph heuristically.
It is applicable to both data- and goal-driven searches and supports a variety of heuristic evaluation functions. Owing to
its generality, best-first search can be used with a variety of heuristics, ranging from subjective estimates of a state's
"goodness" to sophisticated measures based on the probability of a state leading to a goal. This algorithm will continue to
provide a basis for examining the behaviour of heuristic search.
p.129 The first step to solving a problem is to form a goal based on the current situation. As well as formulating
a goal, the factors that affect the desirability of different ways to achieving the goal have to be identified (Luger 2001).
After goal formulation comes problem formulation, in which the actions and states to be considered are decided.
Normally, an agent with several immediate options of unknown values can decide what to do by first examining the different
possible sequences of actions that lead to states of known values, and then choosing the best one. This process of looking
for such a sequence is called searching.
p.130 The main work in the area of search strategies is to find the correct search strategy for
the given problem.
