Information Technology Reference
In-Depth Information
5. Agents can be viewed as simple nutshells in certain cases: they only execute actions
predefined by third party, meaning that their behaviour is defined (controlled) by
experimenter. These agents are called ”Hollow Agents”. For example, a human
trader can act through such agents. By definition, ”Hollow Agents” do not have any
artificial intelligence and can be assimilated to human-machine wrappers.
6. Today, there are only few artificial stock markets presenting multi-assets order topic
([8], [5], [9]). ATOM is one of them, that allows us to simulate more advanced
agents behaviour with risk-return management and wealth optimisation utility func-
tions (see section 4).
7. We overcome a weak point of most existing simulators, that include only limit and
market order types (for instance [8], [10]), ATOM contains all spectrum of orders
present in Euronext Stock Exchange.
3
Artificial Market Design Issues
Artificial Stock Markets (here-after ASM) are environments that allow to express all
classical notions used in multi-agent systems. ASM, like any other MAS, are suited
for the study of various emergents phenomena. Using the so-called ”vowels” approach
which agents evolve is the market microstructure, agents behaviours mimic those of
real traders with their own and interaction occurs through orders and prices. Neverthe-
less, if one wants to build an efficient platform, several issues can be identified and must
be precisely and strictly regulated.
3.1
Entities Organization
During an artificial stock market modelling many MAS design principles are employed,
for example, modularity and encapsulation, that suggest dividing the system into differ-
ent sub-organizations. There are different attempts to organize artificial stock markets.
For example,
Agents
-
Rules
and
Securities
in [12] or
Autonomous engine
(platform
core) -
Simulation User Interface
and
Agents
separate binary modules in [13].
The choice of ATOM organisation is resulting from intention to introduce fairly
tractable markets, to be close to NYSE-EURONEXT stock exchange organisation and
to ground modelling on MAS design main concepts. ATOM architecture can be viewed
as a system with interacting components: i)
Agents
, and their behaviours, ii)
Market
is
defined in terms of microstructure and iii)
Bank
reflects intermediaries and monetary
financial institutions iv) the
Artificial Economic World
provides economical indicators.
Depending upon the researcher's targets, the
Artificial Economic World
can be plugged
or not in the simulations.
We link each system entity with the sets of
Responsibilities
in order to cover all
functionality and complexity of real world market. Thus,
Market
is responsible for the
generating of market scenarios and price path, it presents set of constraints, rules, regu-
lations, leading participants to activities.
Agents
may play roles of buyer or seller with
