Question

In many situations, a user interacts with its computer being part of a distributed network of computers. If the users work in such environment where the other users in the system can affect the results of its own actions the analysis requires an adequate HCAI technique. The multi-agent model can become appropriate. See earlier posted paper: Agenit Based Modeling to Inform the Design of Mulriuser Systems, by Yuqing Ren and Robert E. Kraut Task Consider a distributed network of I workstations, located geographically along a straight line. This network represents a group of agents ai n 1 used to process the available information to evaluate a local geophysical parameter, u. Each agent with exception of a and a has no access to local measuring devices; instead, it relies on the readings produced by its neighbors in the following way: namely, taking and averaging the readings of the neighbors at the last access step kand u to produce the new reading u which becomes available far its neighbors for the step k+1. The measuring devices are available only at the ends of the workstations chain. They produce some constant values:U, andU for all measuring steps k The Ul can be very simple. Three text boxes provide the current information abouta u.1 Each workstation is monitored by a user. As soon as one of the neighbors sets a new reading the user is alerted to recalculate the value of u. That becomes available to others and, in its turn, alerts the neighbors. Questions 1. 2. 3. Does this process converges? Attempt to construct a mathematical model of such HCI system. You can start with analyzing the following numerical example:

Answer 1

yder distributed netuork of n-tuork onJ产 PCal pa au 2ch agst thteption de vicy tu A ve raging t reading the step Ktl an

1) procen . т сп しん金

-うし : 3 uto