Dynamics of large autonomous computational systems Export

@inproceedings{Hogg2002Dynamics, abstract = {Distributed large scale computation gives rise to a wide range of behaviors, from the simple to the chaotic. This diversity of behaviors stems from the fact that the agents and programs have incomplete knowledge and imperfect information on the state of the system. We describe an instantiation of such systems based on market mechanisms which provides an interesting example of autonomous control. We also show that when agents choose among several resources, the dynamics of the system can be oscillatory and even chaotic. Furthermore, we describe a mechanism for achieving global stability through local controls.}, author = {Hogg, Tad and Huberman, Bernardo A.}, booktitle = {Proceedings of the Santa Fe Workshop on Collective Cognition}, citeulike-article-id = {3504594}, citeulike-linkout-0 = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.5.8276}, citeulike-linkout-1 = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.5.8276}, comment = {As in economics [3], the use of prices provides a flexible mechanism for allocating resources, with relatively low information requirements: a single price summarizes the current demand for each resource, whether processor time, memory, communication bandwidth, use of a database or control of a particular sensor. While this example motivates the use of market based resource allocation, a study of actual implementations is required to see how large the system must be for its benefits to appear and whether any of the differences between simple computer programs and human agents pose additional problems Thus we have a second reason to consider markets: not only may they be useful for flexible allocation of computational resources among competing tasks, but also the simplicity of the price mechanism could provide help with designing cooperative parallel programs (see Spawn) Presents results of tasks running in spawn. Parallisation is important. However, it can lead to chaos. or oscillation (mainly due to delays) So, reward agents based on their actual performance. This diversity in turn eliminates chaotic behavior through a series of dynamical bifurcations which render chaos a transient phenomenon. This raises the interesting question of the minimal amount of diversity needed in order to have a stable system. useful. read again.}, keywords = {computational-ecosystem, distributed-computing, market-based-control, network}, pages = {295--315}, posted-at = {2008-11-11 23:25:18}, priority = {0}, title = {Dynamics of large autonomous computational systems}, url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.5.8276}, year = {2002} }

TY - CONF ID - Hogg2002Dynamics L3 - citeulike-article-id:3504594 N2 - Distributed large scale computation gives rise to a wide range of behaviors, from the simple to the chaotic. This diversity of behaviors stems from the fact that the agents and programs have incomplete knowledge and imperfect information on the state of the system. We describe an instantiation of such systems based on market mechanisms which provides an interesting example of autonomous control. We also show that when agents choose among several resources, the dynamics of the system can be oscillatory and even chaotic. Furthermore, we describe a mechanism for achieving global stability through local controls. EP - 315 TI - Dynamics of large autonomous computational systems T2 - Proceedings of the Santa Fe Workshop on Collective Cognition SP - 295 KW - computational-ecosystem KW - distributed-computing KW - market-based-control KW - network N1 - As in economics [3], the use of prices provides a flexible mechanism for allocating resources, with relatively low information requirements: a single price summarizes the current demand for each resource, whether processor time, memory, communication bandwidth, use of a database or control of a particular sensor. While this example motivates the use of market based resource allocation, a study of actual implementations is required to see how large the system must be for its benefits to appear and whether any of the differences between simple computer programs and human agents pose additional problems Thus we have a second reason to consider markets: not only may they be useful for flexible allocation of computational resources among competing tasks, but also the simplicity of the price mechanism could provide help with designing cooperative parallel programs (see Spawn) Presents results of tasks running in spawn. Parallisation is important. However, it can lead to chaos. or oscillation (mainly due to delays) So, reward agents based on their actual performance. This diversity in turn eliminates chaotic behavior through a series of dynamical bifurcations which render chaos a transient phenomenon. This raises the interesting question of the minimal amount of diversity needed in order to have a stable system. useful. read again. AU - Hogg, Tad AU - Huberman, Bernardo PY - 2002/// UR - http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.5.8276 ER -