contact IGD : email
retour IGD
© IGD 2010
Copyright iegd.institut.online.fr - 2010, tous droits réservés

B-ADSc® : the decisional analysis of complex systems
Flexible workshop: about the concept of flexibility
Automation seen as delegation
Techniques de l'Ingénieur - Analyse décisionnelle
Bibliothèque de l'Ecole des Mines de Paris
The decisional analysis of complex systems (otherwise called decisional engineering) consists of designing, analyzing and optimizing an organization so that know-how and the processes employed can be expressed with the most effectiveness and in total convergence of goals with the finality of the unit. Taking account of the assets of the operations research (new process, optimization of the choices, the experience feedback), technological research (creation of resources and new means) and evolutions of the operational context, the decisional analysis endeavors to put in coherence :
Components of an organization structuring factors
  - actors in their activities, divisions of the roles, value systems, information systems
Dynamic components
  - actors' behaviors, wanted, possible, dysfunctions
Methods of evolution
  - changes: spontaneous, stimulated, controlled and under control.

The acceleration of the competition, the increase in the technical possibilities, the modification of the economic context and the eruption in the organizations of intelligent machines acting at the sides of men, is increasing a lot the complexity of the organizations. The architecture of the data-processing solutions then tends to be decentralized for better adapting to the life of the company.
Sat on theoretical bases, decisional analysis brings a systemic answer, pragmatic and formalized in a method :
Bucki - Decisional Analysis of the complex Systems (B-ADSc). Its originator is a Polish engineer, Janusz Bucki, Doctor in mathematics. The applications carried out to date with the assistance of this method, for the account of companies, relate to the engineering of the organizations, the computerization and the automation of the industrial and administrative processes.

The delegation of the decisions, according to the B-ADSc method, is the structuring factor of the organizations. The delegations lead to setting up new activities controlled by men or machines, each one of them constituting an autonomous center of decisions. To be effective, such delegations must be accompanied by corresponding controls. The “delegating” role of the activity moves to become dependent on more general decisions. Following a delegation, the number of decisions to be taken in an organization increases overall, just like the effort required for their development.
The delegation of the decisions can be related on the creation of new activities and the extension of the information system.
Taking into account the cost which results from it, all new delegation requires for validation a value engineering.
This analysis, led to evaluate the potential profits, follows three directions :
Improvement of operation - elaborate more relevant decisions on the level of the system of delegation -
  suppression of nona quality
Increase in the productivity by a reduction in the time of decision-making compared to the rhythm of the process -
  more treated volumes
Increased availability of the actors operating in the roadbases - thanks to the substitution of the decisions usually
  delegated by more total and less frequent decisions (releasing a potential which can be employed again in addition).

By difference of the traditional approaches, B-ADSc describes as complex any organization in which various autonomous actors are likely to contribute to the satisfaction of a common purpose.
By this way, the reactivity, flexibility, efficiency, the reliability and more generally the control of the systems and organizations are obtained easily with tools guaranteeing the treatment of :
  • the man and his behavior
  • the computer and its logic
  • the organization and its Information system directed piloting of the operations.

In reference to Cybernetics, B-ADSc regards as autonomous any entity defined by the following criteria:
  - integrity, not deterioration by the entourage
  - internal structure well defined
  - aptitude for the car determination or the car regulation towards internal or external stresses
  - capacity to articulate itself with the entourage
  - metabolic” independence, detention of its own processor or proper piloting (capacity to act by oneself).

An organization, i.e., any system in which several autonomous actors (men or machines) are likely to contribute to the satisfaction of a common purpose, is characterized by :
  • the participation of the actors, whom it acts men or machines, likely to pursue their own goals while remaining in
    convergence of goals with the common purpose,
  • the autonomy of the actors, taking into account the distribution of the roles and their know-how, in the decision-making
    relating to the stock management which is accessible for them,
  • the pooling of the means and the capacities whose efficiency depends on the existence on procedures explicit and
    implicit rules of assignment and division in agreement with the continued purpose.
  • evolutionarity being able to pass by :
    - redefinition of the common purpose,
    - repositioning of the individual objectives,
    - development or the acquisition of new know-how,
    - creation or acquisition of new means,
    - redeployment of the actors,
  • the opening :
    - capacity to integrate other organizations or other actors,
    - capacity to be integrated in other organizations.

B-ADSc associates complexity the car organization of the autonomous actors.
In addition, the perception of a chaotic evolution could be related to the coexistence, in a universe, of complex organizations from which the finalities are different, even concurrent. Chaos then reflects the interference, on the shared means, of the actions resulting from the various organizations in the absence of a context more general, therefore of a federator purpose. This situation of chaos is distinguished from that relating to the possible dysfunctions of an organization, themselves which had, for example, with a failing communication between the actors or with inadequate know-how.
Due to a new absorbing algebra of activities which constitutes the formal base of it, the B-ADSc method applied to the information processing systems leads in a natural way to the design and the programming (decisional programming) of the Massively parallel systems. Indeed, the vulgarizing of this technology could not be done without an easy and transmissible technique of programming of the massively parallel computers.
The generalization suggested of the logic of the 1st and 2nd order with the logic of the 3rd order opens new prospects in the field for the artificial Intelligence, in particular with regard to the car training or the capacity of the computers to create new conjectures. That makes it possible for example to equip a robot with the aptitude in a new situation to build in an autonomous way a relevant behavior (capacity to face a case not envisaged in advance).