ON
THE COMPUTATIONAL POWER OF ADAPTIVE
SYSTEMS
C.
Cotta, E. Alba, J. M. Troya
Abstract. Recent research has demonstrated that no search algorithm is better than any other one when performance is averaged over all possible discrete problems. Hybridization (incorporation of problem-knowledge) is required to produce adequate problem-specific algorithms. This work explores the power of hybridization in the context of evolutionary algorithms. For this purpose, a framework for describing adaptive systems is presented. It is shown that, when hybridized, adaptive techniques are computationally complete systems with Turing capabilities. Moreover, evolutionary algorithms can be regarded as a kind of nondeterministic Turing machines.
Karhunen-Loève
Transform: An Exercise in Simple
Image-Processing
Parallel Pipelines
M.
Fleury, A. C. Downton, A. F. Clark
Abstract.
Practical
parallelizations of multi-phased low-level image-processing algorithms may
require working in batch mode. The features of a new processing model, employing
a pipeline of processor farms, are described. A simple exemplar, the Karhunen-Loève
transform, is prototyped on a network of processors running a real-time
operating system. The design trade-offs for this and similar algorithms are
indicated. In the manner of co-design, eventual implementation on large- and
fine-grained hardware is considered. The chosen exemplar is shown to have some features, such as
strict sequencing and unbalanced processing phases, which militate against a
comfortable implementation.
Minimum
2-terminal routing in 2-jump circulant graphs
B.
Robič, J. Žerovnik
Abstract.
A 2-jump circulant is an undirected graph whose nodes are integers 0,1, …,
n - 1 and each node u
is adjacent to four nodes u ± h1(mod
n), u
± h2(mod n), where O <
h1 < h2 <
n.
An algorithm for routing a packet along the shortest path between a pair of
processors in 2-jump circulant network is given. The algorithm requires Q(d)
time for preprocessing, and l routing steps, where d
is the diameter of the graph and l = O(d)
is the distance between the two processors.
Group
know-how
René
Pázman
Abstract.
A useful tool for reasoning about multiagent systems (MAS) is a MAS theory,
which also helps with the design of multiagent systems. Most of MAS theories are
exclusively or partially based on a logic formalisation, which allows to define
various properties of agents, such as intentions or knowledge. Most of them deal
with properties of individual agents, but properties of groups of agents are
becoming equally important in some applications too.
This work aims at defining the notion of group know-how within the
framework based on qualitative temporal logic, and tries to show how this
formalisation can be used in the design process of real multiagent systems. The
definition of group know-how is so that to satisfy the success theorem - i.e. if
a rational group knows how to achieve a goal and wants (intends) to achieve it,
it will achieve it in reality.
Theory
of local register allocation for prolog clauses II
M.
Fico
Abstract. When a relative order of an abstract process and of a code rendering is swapped during compilation, then the abstract process relates to compiled clause and not to its abstract code. The proposed algorithms of the optimal register allocation warrant a correspondence between a clause representation and parameters of its abstract code, regardless of the type of occurrence of temporary variables. The feasible optimal abstract code with minimal length and minimal number of registers is an exact defined surroundings for all subsequent optimizing procedures of the Prolog compiler.