Jonathan T. Moore. Safe and efficient active packets. Technical Report MS-CIS-99-24, DSL, University of Pennsylvania, October 1999.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....uses several restrictions to limit resource utilization of active packets in the network. These restrictions consist of a resource bound derived from the time to live (TTL) field and the fact that SNAP programs use bandwidth, CPU, and memory resources in linear proportion to the packet s length [13]. Hence, byte code instructions must execute in a constant and predictable time frame. The resource bound as defined in SNAP turns out to be difficult to respect because byte code instructions may differ significantly in terms of their execution time [14] In addition, the limitation that only ....

....( denotes the packet length) the payload and memory section can be accounted for the packet s resource budget in the same way as code sections. The idea is to use this extra budget for program loops. We define a packet as using the notation borrowed from [13], where is the network part of the resource bound, the full program code, an arbitrary sequence of instructions within , and the memory size containing heap and stack of the packet. The no For simplicity the notation of a packet does not contain all parts of a packet (e.g. ....

[Article contains additional citation context not shown here]

Jonathan T. Moore. Safe and efficient active packets. Technical Report MS-CIS-99-24, DSL, University of Pennsylvania,, October 1999.

.... through the space) 42, 27, 10] ffl to execute parametrized scripts or plan complex task sequences at the higher level (for example, choosing which room to search in order to locate an object or another agent, or outlining the primary steps that must be followed to perform a particular task) [35, 4]. The architecture is built on Parallel Transition Networks PaT Nets [3] Nodes represent executable processes, edges contain conditions which when true cause transitions to another node (process) and a combination of message passing and global memory provide coordination and synchronization ....

M. Moore, C. Geib, and B. Reich. Planning and terrain reasoning. Technical Report MS-CIS-94-63, CIS, University of Pennsylvania, Philadelphia, PA, 1994.

....to lift it, and then carry it to the other location. It is not obvious whether the lifting action is still part of achieving the precondition, or already part of the body of the action. The choice of not having preconditions has been motivated elsewhere in AnimNL with more extensive arguments [Geib, 1992]. Summarizing, Geib s main argument is that Header [CAUSE( agent] i ; GO Sp (j; k) FROM(m) TO(l) # k Body [GO Sp ( i, TO(m) fl 1 [CAUSE(i, GO Ctrl (j, TO( AT(i) fl 2 GO Sp (i; k) WITH(j) # fl 3 Annotations fl 1 enables fl 2 fl 2 enables fl 3 ....

....results. One goal of AnimNL s planner, in elaborating the initial Plan Graph created in response to the specified instructions, is to use bounded look ahead so as to make current goal decomposition decisions based on the curent state of the world and future intentions for further details see [Geib, 1992]. Planning in Anticipation of Acquiring Knowledge AnimNL assumes that an agent cannot have up to date knowledge of any part of its environment that is outside its direct perception. An AnimNL agent may know what nonvisible parts of its environment were like, when it perceived them earlier, and ....

Christopher Geib. Intentions in Means/End Planning. Technical Report MSCIS -92-73, University of Pennsylvania, 1992.

.... through the space) 37, 24, 7] ffl to execute parametrized scripts or plan complex task sequences at the higher level (for example, choosing which room to search in order to locate an object or another agent, or outlining the primary steps that must be followed to perform a particular task) [31, 4]. The architecture is built on Parallel Transition Networks (PaT Nets) 3] Nodes represent executable processes, edges contain conditions which when true cause transitions to another node (process) and a combination of message passing and global memory provide coordination and synchronization ....

M. Moore, C. Geib, and B. Reich. Planning and terrain reasoning. Technical Report MS-CIS-94-63, CIS, University of Pennsylvania, Philadelphia, PA, 1994.

....plan graph. These processes include parsing, interpretation and plan inference [18, 19, 20, 21] ffl A module consisting of a high level incremental planner and two specialized processes able to adapt highly parameterized plans for search and for object manipulation to the exact situation at hand [26, 27, 28, 36, 41]. ffl A simulator that coordinates motion directives and perceptual requests from the planning components with ones corresponding to environmental responses, and schedules their performance. An agenda allows multiple behaviors to be carried out in parallel, and other behaviors to be initiated and ....

....that the red file folder is in Fred s office. What is especially interesting is the strength of this expectation: a cooperative agent will look around, if an object isn t where they expect it to be until they find it. This has led Moore to develop flexible procedures he calls search plans [41] following [39] that can be used to guide an agent in grounding both definite and indefinite referring expressions. Moore s search plans are able to incorporate expectations about the context in which a referring expression will receive its intended grounding, to limit search. In AnimNL, Di ....

Moore, M.B. Search Plans. PhD Dissertation proposal, Department of Computer and Information Science, University of Pennsylvania, May 1993. (Technical report MS-CIS-93-55).

....plan graph. These processes include parsing, interpretation and plan inference [18, 19, 20, 21] ffl A module consisting of a high level incremental planner and two specialized processes able to adapt highly parameterized plans for search and for object manipulation to the exact situation at hand [26, 27, 28, 36, 41]. ffl A simulator that coordinates motion directives and perceptual requests from the planning components with ones corresponding to environmental responses, and schedules their performance. An agenda allows multiple behaviors to be carried out in parallel, and other behaviors to be initiated and ....

....using intentions and limited simulation to perform action selection means the possibility of action failure must be allowed for. The current version of AnimNL s planner, which eliminates preconditions in favor of situated reasoning about the effects of actions, is described in more detail in [26, 27]. 4 Expectations from Instructions Here we address the role of instructions in raising expectations that complement an agent s current perceptions in influencing its behavior. Expectations can lead to further perceptual activity not just observation but also activities that enable observation. ....

Geib, C. Intentions in Means/End Planning. Technical Report MS-CIS-92-73, Dept. Computer & Information Science, University of Pennsylvania, 1992.

No context found.

Jonathan T. Moore. Safe and efficient active packets. Technical Report MS-CIS-99-24, DSL, University of Pennsylvania, October 1999.

No context found.

J. T. Moore. Safe and e#cient active packets. Technical Report MS-CIS-99-24, University of Pennsylvania, October 1999.

No context found.

J. T. Moore. Safe and e#cient active packets. Technical Report MS-CIS-99-24, University of Pennsylvania, October 1999.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC