Aaron Kershenbaum

Centralized teleprocessing network design

The problem considered is that of finding an optimal (minimum cost) design for a centralized processing network given a set of locations, traffic magnitudes between these locations, and a single common source or destination. Several heuristics, which are efficient (in terms of their execution time and memory requirements on a digital computer) and which produce seemingly good results, have already been developed and are currently accepted techniques. Some work has also been done on finding optimal solutions to this problem both as a design tool and as a means of verifying the effectiveness of proposed heuristics. We focus on this latter area. Currently known techniques for the optimal solution of this problem via integer programming have fallen short of the desired objectives as they require too much memory and running time to be able to treat problems of realistic size and complexity. We develop an improved technique which is capable of handling more realistic problems.

Computing capacitated minimal spanning trees efficiently

An analysis is made of the computational complexity of a class of heuristic algorithms for the solution of the minimal spanning tree problem subject to a restriction on the maximum number (or weight) of nodes in any subtree rooted at a distinguished node. This is of particular interest in designing networks with branch capacity restrictions. The algorithm is a modification of Kruskals Algorithm where weights are assigned to the nodes and then used, along with the arc lengths, to select the order in which arcs are considered for inclusion in the spanning tree. Considerations in the efficient implementation of such algorithms are examined and several heuristics for assigning node weights are compared.

A Distributed Evolutionary Algorithm for Reorganizing Network Communications

The Distributed Evolutionary Algorithm (DEA) presented in this paper produces efficient TDMA schedules for communication in an environment of changing network topology by passing topological and traffic information among nodes communicating on the schedule. As information is gained in this way more nodes are added to the communicating group until the entire network is communicating on a common schedule. The key feature of the DEA presented is that the reorganization phase is fast, requiring a number of TDMA slots only on the order of the number of nodes. A secondary feature is that some network communication can take place, and thus some of the traffic requirements can be satisfied, even while the network is being reorganized. Numerical results are presented for a number of randomly generated networks which show how the algorithm performs.

An Algorithm for Evaluation of Throughput in Multihop Packet Radio Networks with Complex Topologies

The problem of analyzing the thoughput of packet radio networks with realistic topologies is considered. We present an algorithm for the solution of this problem and show that both the memory requirements and running time of this algorithm in practice grow polynomially with the size of the problem. Although in theory both can grow exponentially in the worst case, we offer computational experience with the procedure and show that for realistic topologies where connectivity is related to distance, the rate of growth is quadratic in the number of links. Even for regular grids, which are pathological in their symmetry, the rate of growth is only cubic in the number of links. We thus conclude that the procedure is effective for realistic topologies with up to several hundred nodes.

Scheduling multihop CDMA networks in the presence of secondary conflicts

This paper presents an algorithm for producing near-optimal conflict-free schedules for networks operating under code division multiple access (CDMA). A procedure for finding a lower bound on the length of such schedules is also presented. The presence of both primary and secondary conflicts (due to imperfectly orthogonal CDMA codes) are accounted for by these algorithms. The complexity of both algorithms is analyzed and computational experience with both procedures is presented. Using the lower bound, it is shown that the heuristic is effective. The complexity analysis demonstrates that it is efficient enough to use in networks of realistic size, even when the schedules must be produced in real time.

A note on finding shortest path trees

Two shortest path algorithms are compared and it is shown that, while one outperforms the other in practice, the formers running time is exponential in the worst case while the latters is polynomial. A procedure which constructs such worst case examples is given.

Minimal Length Test Sequences for Protocol Conformance

A new procedure is described to produce minimal length test sequences for protocol conformance. The procedure is applicable to testing based on Unique Input/Output (UIO) sequences, Distinguished Sequences, and other testing methodologies. The procedure efficiently exploits the overlapping among testing segments to the maximum extent.

Network management and control of protocols

The authors describe some of the special requirements placed on the protocols that must function in a multidomain environment. Taking these requirements into consideration, they describe a novel protocol structure called a bipartite protocol, which is based on the use of tokens. They show that the use of bipartite protocols makes it possible to perform a number of functions, namely to systematically explore subgraphs of the global state diagram representing the protocol, with predetermined numbers of messages in the channel queues, thereby facilitating the analysis of protocols; to generate tightly synchronized protocols or sections of protocols, thereby controlling the build up of queues in the channels; and to implement sections of network control systems using the tokens in the bipartite protocol structure. Examples are given for X.75, illustrating both the analysis of protocols and the control of channel contents.<<ETX>>

A Dictionary Matching Algorithm Fast on the Average for Terms of Varying Length

We examine the exact dictionary matching problem with dynamic text and static terms and propose a simple but efficient algorithm with sublinear (in size of text) average performance for a wide range of practical problems. The algorithm is based on the Commentz-Walter-Horspool algorithm (CWH), presented by Baeza-Yates and Re‘gnier [101. Typically, our refinement will prune out more than 30% of characters scanned by CWH, when searching for all occurrences of tags, which are of varying lengths and members of a set of moderate size, in natural language text. This problem arises frequently in practice in scanning text downloaded from the internet, and accounts for a major portion of the preprocessing time associated with indexing such text for later retrieval. Our approach, which we refer to as layering, keeps track of an upper bound on the maximal length of potential term prefixes ending at each given position in the text. This information is then used to mask out some of the terms and filter out unnecessary character comparisons during the search. A practical implementation is described, which increases the size of the existing data structures as well as the preprocessing cost only by a factor of the size of the longest term in the set.

An Application of Fuzzy Multi-Criteria Decision Making for Topological Design of Large Networks

In this paper, techniques of fuzzy nultiple attribute decision making are applied to the design and analysis of large network topologies. A theoretical approach is presented that is computationally efficient, and broadly applicable to a wide class of large network design problems. This approach was developed to allow rapid analysis of performance trade-offs inherent in a broad range of topological network structures, and to assess and identify the sensitivity of design selection to changes in the decision criteria.