Simulation and Reality
Fig. 2
The labels on the arrows describe the flows of information or actions among the various sources and repositories of information. The operations sites are not staffed and communicate by telemetry with the operations control centre. Central Systems and Work Planning & Scheduling are computer systems. The rest are of the boxes represent people or the persons grouped as departments or sections
This notion of the "statistical signature" as defined by Arthur et. al. [14] seems to be a statement about the visual appearance of a line chart. Though a useful notion, it is not sufficiently well defined to provide as clear a verification of a model as was obtained in the models of the previous section. In this section, a model is reported in which the output is characterised by a plausible series of outcomes conforming to the statistical signature associated with observations of the empirical referent of the model. The model also contains an explicit and well validated representation of agent cognition together with an accurate description of the relevant information systems and organizational structure of an actual company. The model is concerned with the systems and procedures for responding to critical incidents in the water and sewage services industry. Critical incidents include those which are likely to interrupt the provision of these services to the public or which will cause environmental damage or pollution but are containable with the existing assets, systems and procedures normally available to managers of the company.
The systems and organizational structure of the company as they relate to critical incidents is depicted in fig. 2. The cognitive agents in the model are the network controllers and the operations control centre. The model was implemented in SDML with the container structure depicted in fig. 3. The model cycles over days and, within each day, 18 task cycles. At the end of every six task cycles, the network controller changes although the same three controllers are active in the same rotation each day.
Fig. 3
The agent of type PhysicalWorld decides which, if any, primary events will occur spontaneously according to the specified probabilities and, if any such event does occur, assigns it to an operating site at random. If there are already events occurring at an operating site, the PhysicalWorld propagates consequential events at random from the specified probabilities and assigns them to the appropriate operations site. After the PhysicalWorld determines the state of the world, all of the agents in the social world fire their rules in parallel. They cycle over the time period elaborationCycle which is a step in elaborating mental models. After each cognitive of artificially intelligent agent determines any mental models it requires and takes such actions as are implied by those mental models, if is finished for that task cycle
The causes of specific events are not in practice known to the individuals involved in critical incident management until the manifestations of the incident have been observed and the situation analysed. Even then, they might make the wrong diagnosis. For these reasons, it would not be appropriate to post the causes of a particular incident to a database accessible by all agents. Consequently, the relevant clauses are asserted privately by the PhysicalWorld to its own databases and the fact of the occurrence of the event is asserted to the database of the operation site at which the event is to occur. This assertion is achieved by the explicit addressing of the clause eventOccuring event where event is actually fire, pumpFailure, contaminationIncident, or the like. Once one such event has been allocated to the operating site, then with the appropriate probability all of the consequential events and their consequential events, etc. are also allocated at random to that site. In subsequent time frames, secondary consequences are allocated to that site with the specified probabilities in every event cycle while events with those secondary consequences continue. Events, once allocated, remain a feature of the site until they are remedied, if there are remedies, and the events which gave rise to them have been eliminated.
The operating sites (in practice mystified) recognize two kinds of event: telemetric and publicly observable events. When a site has had a telemetered event asserted to its databases, it sends a message stating that it has that event to the OperationsControlCentre. When a site has a publicly observable event asserted to its databases, it selects at random a percentage of households and asserts the occurrence of that event to their databases. In the simulations reported here, each household of 100 had a 10 per cent probability of being selected to receive such a message. Because the information contained in those assertions refers to actual information which is available selectively, once again explicit addressing of the assertions is appropriate.
The OperationsControlCentre agent forwards the telemetry and public reports to the CentralSystems agent who decides on the actions to be taken. The instructions to take these actions are addressed explicitly to the WorkPlanningAndScheduling agent who allocates the work by addressing the instructions of an agent of type RepairGang or Controller as appropriate. The reports by the repair gangs or controllers are addressed to CentralSystems agent. Repairs take the form of actions asserted by the repair gang to the operating site and then read from the operating site's databases by the PhysicalWorld instance.
The cognitive behaviour in this model is by the instances of type Controller, and the workPlanningAndfScheduling, OperationControl and CentralSystems instances. These agents learn by specifying and testing models which are held on their respective databases as private clauses -- i.e. clauses which can be asserted by an agent only to its own database and read only by that agent.
Generated with CERN WebMaker
Simulation and Reality - 20 MAY 98
[Next] [Previous] [Top] [Contents]