Ramalingam, G. and Reps, T., "On the computational complexity of incremental algorithms," TR-1033, Computer Sciences Department, University of Wisconsin, Madison, WI (August 1991).  Home/Search   Document Not in Database   Summary   Related Articles   Check

....this notion precise in a way which is independent of particular incremental algorithms is not always straightforward. Two early approaches along these lines are Goodwin [6, 7] reason maintenance) and Reps [13] language based editing) More recently, Alpern et al. 1] and Ramalingam and Reps [12] have provided a framework for analysing incremental algorithms, in which the basic measure used is the sum of the sizes of the changes in the input and output. This framework presupposes a problem instance (a representation of the current input) P and a solution S (the current output) When P ....

G. Ramalingam and Thomas W. Reps. On the Computational Complexity of Incremental Algorithms. Technical Report 1033, Computer Sciences Department, University of Wisconsin, Madison, Wisconsin, 1991.

....domain that also have good knowledge about algorithms. As the word ad hoc suggests, these algorithms themselves do not provide a general systematic approach for obtaining them. Complexity theoretic issues of incremental algorithms have received much attention in the past few years [BPR90,Mil91,RR91,Ber92,SVT93,Ram93,MSVT94] Although these works do not provide methods for obtaining incremental algorithms, they can assist the study of general approaches to incremental computation by establishing certain theoretical bounds. Research in this area is yet in an early stage. 2.2 Incremental ....

G. Ramalingam and Thomas Reps. On the computational complexity of incremental algorithms. Technical Report TR-1033, Computer Sciences Department, University of Wisconsin, Madison, Wisconsin, August 1991.

....results in a considerable speedup and will be used throughout this paper without an explicit mention. We require that the incremental algorithms in this paper take less time than the corresponding non incremental algorithm. This criterion differs from the work done by Ramalingam et al.[11], which studied incremental algorithms for certain graph problems. This paper defined incremental algorithms as those algorithms, whose time complexity could be written as a function of the change to the system alone, where the change or D could be written as the sum of the change in the input and ....

G. Ramalingam and T. Reps, "On the Computational Complexity of Incremental Algorithms," Tech. Rep. TR 1033, University of Wisconsion, Madison, University of Wisconsion, Madison, 1991.

....algorithm starts the incremental update with the vertices that have been modified by the update and proceeds by following the pointers in the storage blocks of the vertices. The algorithm may keep global information only about vertices it has visited during the current update. Ramalingam and Reps [66] showed that there is no bounded incremental algorithm in the locally persistent model for the single source reachability (SSREACHABILITY) problem, which is defined as follows: given a directed graph (V; E) and a fixed vertex s 2 V , find for every vertex v 2 V whether v is reachable from s. We ....

G. Ramalingam and T. Reps. On the computational complexity of incremental algorithms. Technical Report TR-1033, Computer Sciences Dept., University of Wisconsin-Madison, 1991.

....= 9 y T (y; x)Q(y) Q(x) R(x) LFP (f; I(x) return R(x) Unfortunately, this algorithm is not incremental, and if the designer modifies the system, the reachable states have to be computed from the beginning. Incremental algorithms for certain graph problems were examined by Ramalingam et al. [11]. They defined incremental algorithms as algorithms, whose complexity was a function of the size of the change to the system alone. If Delta = Delta input Delta output , and Delta input is the change in the input to the program, Delta output is the change that results in the output, then ....

G. Ramalingam and T. Reps, "On the Computational Complexity of Incremental Algorithms," Tech. Rep. TR 1033, University of Wisconsion, Madison, University of Wisconsion, Madison, 1991.

....to meet new specifications is very valuable for many tasks. Such reuse can provide substantial computational advantages by avoiding repetition of redundant work. Consequently, incremental modification has emerged as an important research issue in constraint solving and many other areas [142]. Constraint languages are not yet able to cope with a dynamically changing environment, for instance when the user wants to modify some data, add a new input or relax a previous constraint. Limited results has been obtained for non reactive CLP systems [85] but they need to be extended and ....

G. Ramalingam and T. Reps. On the computational complexity of incremental algorithms. Computer Science Technical Report 1033, University of Wisconsin at Madison, August 1991.

....prove that in the worst case no incremental algorithm that handles edge deletions can be faster than an algorithm that solves the problem from scratch [EG85] This work inspired the developments described in Chapter 4. Work begun by Alpern et al. AHR 90] and extended by Reps and Ramalingam [RR91] gives lower bounds based on ffi analysis, within a limited model of computation called local persistence; their work is described and extended further in Chapter 5. The contributions of this dissertation are as follows: ffl to define a framework in which incremental lower bounds can be derived; ....

....algorithms are based, to programs whose loop nesting level is bounded by a constant, we can improve the worst case bound from O(n 2 ) to O(n) RP88] Similar results may be achievable with incremental data flow algorithms. 37 Chapter 5 ffi Analysis of Incremental Algorithms 5. 1 Overview In [RR91] Ramalingam and Reps argue that classifying incremental algorithms by their worst case asymptotic performance may be unnecessarily pessimistic in many cases. As an alternative, they suggest reporting the cost of the update as a function of the size of the changes required by the update; this ....

[Article contains additional citation context not shown here]

G. Ramalingam and T. Reps. On the computational complexity of incremental algorithms. Technical Report #1033, Computer Sciences Department, University of Wisconsin--Madison, 1991.

....amount of work absolutely necessary to perform a given incremental problem. It does not take into account updating costs of various internal data structures used by the particular incremental algorithm. changed is also not known a priori; when the update process begins, only j DeltaP j is known [RR91] An incremental algorithm is bounded if the computation of the solution update takes time dependent only on changed, and not on the size of the entire input. Otherwise, an incremental algorithm is unbounded. A problem is said to be incremental if it has a bounded algorithm. Boundedness is not ....

....of the entire input. Otherwise, an incremental algorithm is unbounded. A problem is said to be incremental if it has a bounded algorithm. Boundedness is not the only relevant criterion in the study of incremental computation, and in fact relatively few bounded incremental algorithms are known [RR91] Some examples of research on incrementality are the following. Ramalingam and Reps [RR91] have studied incrementality in general and they order known incremental algorithms into a complexity hierarchy. They emphasize that the sum of the sizes of the changes in the input and in the output is a ....

[Article contains additional citation context not shown here]

G. Ramalingam and Thomas Reps. On the computational complexity of incremental algorithms. Computer Sciences Technical Report #1033, University of Wisconsin-Madison, August 1991.

No context found.

Ramalingam, G. and Reps, T., "On the computational complexity of incremental algorithms," TR-1033, Computer Sciences Department, University of Wisconsin, Madison, WI (August 1991).

No context found.

Ramalingam, G. and Reps, T., "On the computational complexity of incremental algorithms," TR-1033, Computer Sciences Department, University of Wisconsin, Madison, WI (August 1991).

No context found.

Ramalingam, G. and Reps, T., "On the computational complexity of incremental algorithms," TR1033, Computer Sciences Department, University of Wisconsin, Madison, WI (August 1991).

.... Succ (u) do [13] rhs (v) g v (d (x 1 ) d (x k ) 14] if rhs (v) d (v) then [15] AdjustHeap(Heap, v, min(rhs (v) d (v) 16] else [17] if v Heap then Remove v from Heap fi [18] fi [19] od [20] else u is underconsistent [21] d (u) 22] for v (Succ (u) u ) do [23] rhs (v) g v (d (x 1 ) d (x k ) 24] if rhs (v) d (v) then [25] AdjustHeap(Heap, v, min(rhs (v) d (v) 26] else [27] if v Heap then Remove v from Heap fi [28] fi [29] od [30] fi [31] od end postconditions Every vertex in V (G) is consistent ....

.... GlobalHeap fi [15] fi end begin [16] GlobalHeap : 17] for every production p P do [18] recomputeProductionValue(p) 19] od [20] while GlobalHeap do [21] Select and remove from GlobalHeap a non terminal X with minimum key value [22] if key (X) d (X) then X is overconsistent [23] d (X) key (X) 24] SP (X) p p is a production for X such that value (p) d (X) 25] Heap (X) 26] for every production p with X on the right hand side do [27] recomputeProductionValue(p) 28] od [29] else X is underconsistent [30] d (X) 31] SP (X) p p is a ....

[Article contains additional citation context not shown here]

Ramalingam, G. and Reps, T., "On the computational complexity of incremental algorithms," TR-1033, Computer Sciences Department, University of Wisconsin, Madison, WI (August 1991).

No context found.

G. Ramalingam and Reps T. On the computational complexity of incremental algorithms. Technical report, University of Wisconsin-Madison, August 1991.

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