These are my most significant published works. They are classified in:
Temporal Constraint Satisfaction is an
information technology useful for representing and answering queries about
the times of events and the temporal relations between them.
Information is represented as a Constraint Satisfaction Problem (CSP)
where variables denote event times
constraints represent the possible temporal relations between them.
The main tasks are two:
(i) deciding consistency,
(ii) answering queries about scenarios that satisfy all constraints.
This paper overviews results on several classes of Temporal CSPs:
qualitative interval, qualitative point, metric point, and some of their combinations.
Research has progressed along three lines:
(i) identifying tractable subclasses,
(ii) developing exact search algorithms, and
(iii) developing polynomial-time approximation algorithms.
Most available techniques are based on two principles:
(i) enforcing local consistency (e.g. path-consistency),
(ii) enhancing naive backtracking search.
Despite the ubiquity of time and temporal references in legal texts,
their formalization has often been disregarded or addressed in an
Abstract: The approach to temporal reasoning which has proven most popular in AI is the reified approach. In this approach, one introduces names for events and states and uses special predicates to assert that an event or state occurs or holds at a particular time. However, recently the reified approach has come under attack, both on technical and on ontological grounds. Thus, it has been claimed that at least some reified temporal logics do not give one more expressive power than provided by alternative approaches. Moreover, it has been argued that the reification of event and state types in reified temporal logics, rather than event and state tokens, makes the ontology more complicated than necessary.
In this paper, we present a new reified temporal logic, called TRL,
which we believe avoids most of these objections. It is based on the
idea of reifying event tokens instead of event types. However, unlike
other such attempts, our logic contains ``meaningful'' names for
event tokens, thus allowing us to quantify over all event tokens that
meet a certain criterion. The resulting logic is more expressive than
alternative approaches. Moreover, it avoids the ontologically
objectionable reification of event types, while staying within classical
Temporal Constraint Networks are a well-defined, natural and
efficient formalism for representing temporal knowledge based on metric
They support the representation of both metric and some qualitative temporal
relations and are provided with efficient algorithms based on CSP techniques.
Recently, a generalization based on fuzzy sets has been proposed
in order to cope with vagueness in temporal relations.
In this paper
we generalize some earlier definitions for Fuzzy Temporal Constraint Networks,
we identify and define ``interesting'' queries in a fuzzy temporal
and explore a method for efficiently computing them in a specific case.
Further analysis of some measures on possibility distributions turns out to be
fundamental in order to precisely determine some of these queries.
We discuss the advantages and shortcomings of various choices
and propose specific alternatives which satisfactorily avoid the problems of
The results presented in this paper can be useful
in the design of
a system for temporal reasoning under uncertainty.
For instance, we have applied them to define a
possibilistic temporal logic
in Possibilistic Temporal Reasoning based on Fuzzy Temporal Constraints
where approximate and temporal representation and reasoning are
The notion of time is ubiquitous in any activity that requires
intelligence. In particular, several important notions like change,
causality, action are described in terms of time.
Therefore, the representation of time and reasoning about time is of crucial
importance for many Artificial Intelligence systems.
Specifically during the last 10 years, it has been attracting the attention
of many AI researchers. In this survey, the results of this work are analysed.
Firstly, Temporal Reasoning is defined.
Then, the most important representational issues which determine a
Temporal Reasoning approach are introduced: the logical form on which
the approach is based, the ontology (the units taken as primitives,
the temporal relations, the algorithms that have been developed, ...)
and the concepts related with reasoning about action (the representation of
change, causality, action, ...).
For each issue the different choices in the literature are discussed.
We present empirical evidence that the
effort required to solve CSPs
randomly generated at the 50% satisfiable point and
solved by backtracking based algorithms,
can be approximated
by two standard continuous probability distributions functions.
Solvable problems are quite well modelled by the Weibull distribution function,
and unsolvable problems by the inverse Gaussian distribution.
In this paper we propose a propositional temporal language based on fuzzy
temporal constraints which turns out to be expressive enough for domains -like
many coming from medicine- where knowledge is of propositional
nature and an explicit handling of time, imprecision and uncertainty are
The language is provided with a natural possibilistic semantics to
account for the uncertainty issued by the fuzziness of temporal constraints.
We also present an inference system based on specific rules dealing with the
temporal constraints and a general fuzzy modus ponens rule whereby
behaviour is shown to be sound.
The analysis of the different choices as fuzzy operators leads us to identify the
well-known Lukasiewicz implication as very appropriate to define the
notion of possibilistic entailment, an essential element of our inference system.
Most of AI research on temporal reasoning has been devoted to either
exploring constraint-based temporal deduction techniques
or investigating diverse logics extended with time.
Nevertheless the formal study of deductive systems for such
temporal extended logics has received little attention.
This paper presents a general framework for temporal reasoning in
knowledge-based systems resulting from embedding a temporal reasoner
into a general calculus.
From a representational point of view it is based on the notions of temporal
token and temporal constraint.
The logic is formally defined as a particular
many-sorted predicate calculus and provided with an complete and sound
inference system composed of non-temporal and temporal inference rules.
Moreover, a deductive procedure is presented and analyzed.
It is a general forward chaining algorithm for which
soundness and completeness are guaranteed.
Abstract: Time is fundamental in representing and reasoning about changing domains. A proper temporal representation requires characterizing two notions: (1) time> itself, and (2) temporal incidence, i.e. the domain-independent properties for the truth-value of fluents and events throughout time. There are some problematic issues such as the expression of instantaneous events and instantaneous holding of fluents, the specification of the properties for the temporal holding of fluents and the Dividing Instant Problem.
This paper presents a theory of time and temporal incidence which
is more natural than its predecessors and
satisfactorily addresses the issues above.
Our theory of time, called IP, is based on having instants and periods at equal level.
We define a theory of temporal incidence upon it
whose main original feature is the distinction between continuous and discrete fluents.
Instants have been criticised as temporal primitive for common-sense reasoning
-for being just abstract entities not concerned with
common-sense- and semantical - the Divided Instant Problem (DIP)-
arguments. Therefore, period-based theories have received special attention.
In this paper we provide arguments for incorporating instants to the time
We present an axiomatization of time based on instants and periods
characterize all the models of the theory, and
explore its relations with other Period- and Instant-Period-based theories
appeared in the AI community.
Finally, we discuss its suitability for supporting temporal knowledge
representation and the Divided Instant Problem is revisited.
This page is maintained by Lluís Vila.
Last update: 1 June 99