Journal of the Operations Research Society of Japan
Online ISSN : 2188-8299
Print ISSN : 0453-4514
ISSN-L : 0453-4514
Volume 41, Issue 1
Displaying 1-17 of 17 articles from this issue
  • Article type: Cover
    1998 Volume 41 Issue 1 Pages Cover1-
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    Download PDF (18K)
  • Article type: Appendix
    1998 Volume 41 Issue 1 Pages App1-
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    Download PDF (59K)
  • Konosuke Kawashima, Issei Kino, Naoki Makimoto, Masakiyo Miyazawa, Yos ...
    Article type: Article
    1998 Volume 41 Issue 1 Pages 1-2
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    Download PDF (258K)
  • Shigeo Shioda
    Article type: Article
    1998 Volume 41 Issue 1 Pages 3-20
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    The worst-case performance of an ATM multiplexer, in which each multiplexed connection Sends cells so as to conform to a generic cell-rate algorithm (GCRA), has been examined. Specifically, the concept of large-deviation ordering is introduced to find the "worst" cell-arrival process, i.e., the one yielding the worst performance of the ATM multiplexer. Several interesting results were obtained. First, the cell-loss ratio (CLR) of a bufferless model does not always give the worst CLR when the buffer size per connection is less than c/M, where c is the capacity of the outgoing link per connection, and M is the peak rate of the connection. Second, the cell-delay variation tolerance does not affect the worst-case performance if the buffer size is sufficiently large. Finally, theoretical support was derived for Yamanaka's conjecture that the greedy on-off pattern is not always the worst cell-arrival pattern among those conforming to the dual GCRAs.
    Download PDF (1240K)
  • Shoji Kasahara, Toshiharu Hasegawa
    Article type: Article
    1998 Volume 41 Issue 1 Pages 21-34
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    This paper considers the loss probability of the multiplexing system with the large size of the buffer and the large number of homogeneous ON-OFF sources. We model the multiplexing system as the fluid queueing system with the regulated input fluidt raffic. Each traffic generated by an ON-OFF source (original traffic) is regulated by the leaky bucket and aggregated output traffic (regulated traffic) flows into the buffer of the multiplexer. Using the CDE approximation method, which consists of the loss probability for the bufferless system and that for the system with the large buffer, we explicitly derive the approximated probability of the buffer overflow for the multiplexing system. Our result includes the two important parameters of the leaky bucket; the size of a token pool and the token generation rate. Hence, our result provides the significant indication for designing the input source regulator with the leaky bucket.
    Download PDF (1056K)
  • Shingo Ata, Tetsuya Takine, Masayuki Murata, Hideo Miyahara
    Article type: Article
    1998 Volume 41 Issue 1 Pages 35-52
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    ATM Block Transfer (ABT) is promising for effectively transferring highly bursty data traffic in ATM networks. We develop an approximate analysis for two types of ABT (ABT/DT and ABT/IT) to discuss performance comparisons of ABT/IT and DT. Through numerical examples, we show that ABT/DT is quite sensitive to propagation delays while ABT/IT is not. We next investigate the performance improvement by dynamic bandwidth negotiation, which is applicable to ABT/DT. Bandwidth reduction methods in both of ABT/DT and IT are also examined. Simulation results show that the bandwidth negotiation can improve the performance of ABT/DT, and that in the case of short propagation delays, it outperforms even ABT/IT in terms of throughput. However, it is obtained at the expense of the increased burst transmission times. On the other hand, the bandwidth reduction allows a flexible use of the bandwidth, leading fairly good performance in all parameter regions.
    Download PDF (1547K)
  • Hiroshi Toyoizumi, J. George Shanthikumar, Ronald W. Wolff
    Article type: Article
    1998 Volume 41 Issue 1 Pages 53-67
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    A new stochastic process called block renewal process (or G(m) process) is proposed for modeling auto-correlated arrival processes of queueing systems. This process can be modeled easily by the observable marginal distribution and global auto-correlation index of auto-correlated arrival process. This new process is very simple and tractable, so it is good for both simulation and numerical computation. We also discuss how to obtain analytically the steady-state distribution of queue length of a G(2)/M/1 queue.
    Download PDF (901K)
  • Yutaka Sakai, Yoshitaka Takahashi, Yutaka Takahashi, Toshiharu Hasegaw ...
    Article type: Article
    1998 Volume 41 Issue 1 Pages 68-80
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    We consider a finite-capacity M/G/1/K queueing system with a mixture of the single vacation, close-down, and set-up times; arising out of the switched virtual channel connection (SVCC) mechanism for an IP over ATM network. Applying the supplementary variable approach, we derive the steady-state queue length and waiting time distributions, as well as the set-up occurring rate. Taking the limit of our results as the capacity (K) tends to infinity, we obtain the results for an infinite-capacity queueing system, covering the previous work by Hassan and Atiquzzaman as a special case.
    Download PDF (845K)
  • Ting-jie Lu, Wuyi Yue, Toshiharu Hasegawa
    Article type: Article
    1998 Volume 41 Issue 1 Pages 81-90
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    In this paper, we consider a special mutual overflow system for the analysis of RTNR (Real Time Network Routing) dynamic routing networks. The system has three multiserver service facilities and no waiting room is available. The first-offered traffic to a facility is generated to form an independently and identically distributed Poisson process. If all the servers of a facility are busy, the blocking occurs. Blocked customer overflows to the other facilities and service is required from both of the rest facilities simultaneously. If the service can be provided immediately, then two servers that belong to their respective facility are assigned to the customer and start or finish service at the same time. Otherwise, the overflows will be cleared from the system. For such a system, we describe an algorithm to compute exactly the joint state probabilities by considering the server reservation for first-offered traffic. Following this result, we give a series of steady-state performance measures including blocking probability, carried traffic, utilization of facilities and so on. Numerical examples will also be discussed for understanding of the behavior of load balance mechanism in such a system and with simultaneously to show the effectiveness of the presented performance evaluation method.
    Download PDF (1012K)
  • Atsuhiro Tanaka, Issei Kino
    Article type: Article
    1998 Volume 41 Issue 1 Pages 91-117
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    The dynamlic behavior of programs is studied through LRU (Least-Recently-Used) stack distribution assuming Markovian page references. We propose a new analysis method to calculate steady state probabilities for the LRU stack. We first prove that a Markov chain which describes the behavior of an LRU stack of full-length is lumpable to a Markov chain having a smaller transition matrix than that in the original Markov chain. Based on this lumpability, we propose an algorithm to calculate the invariant vector for the transition matrix of the Markov chain for a full-length LRU stack. We show numerical examples calculated with the proposed algorithm, and evaluate computational efficiency of the proposed method.
    Download PDF (1733K)
  • Kou Fujimoto, Yukio Takahashi, Naoki Makimoto
    Article type: Article
    1998 Volume 41 Issue 1 Pages 118-141
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    This paper is concerned with geometric decay properties of the joint queue length distribution p(n_1,n_2) in two-stage tandem queueing system PH/PH/c_1 → /PH/c_2. We prove that, under some conditions, p(n_1,n_2) 〜 C(n_2)η^<n_1> as n_1 → ∞ and p(n_1,n_2) 〜 C^^-(n_1)η^^-^<n_2> as n_2 → ∞. We also obtain the asymptotic form of state probabilities when n_1 is large or when n_2 is large. These results prove a part of the conjecture of a previous paper [1]. The proof is a direct application of a theorem in [7] which proves geometric decay property of the stationary distribution in a quasi-birth-and-death process with a countable number of phases in each level.
    Download PDF (1393K)
  • Hideaki Takagi, Tsuyoshi Nishi
    Article type: Article
    1998 Volume 41 Issue 1 Pages 142-151
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    Correlation of consecutive interdeparture times characterizes the output process in a queueing system. We present a recursive procedure for calculating the joint distribution of an arbitrary number of consecutive interdeparture times in M/G/1 and M/G/1/K queues. We then obtain explicitly the covariances of nonadjacent interdeparture times, and display the correlation coefficients that reveal the long-range dependence.
    Download PDF (650K)
  • Daiji Horibe, Naoto Miyoshi
    Article type: Article
    1998 Volume 41 Issue 1 Pages 152-165
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    We investigate the sensitivity analysis for a discrete-time queueing system using perturbation analysis. In discrete-time queues, not only the sample performance function but also the sampled event lifetimes are essentially discontinuous in some system parameter, and therefore the well-known infinitesimal perturbation analysis (IPA) technique fails to apply to even a simple single-server queue. We here apply an idea similar to the smoothed perturbation analysis for probabilistic routing problem (or equivalently, negative customer rare perturbation analysis), and under the stationarity and ergodicity assumptions, we obtain the strongly consistent estimator. Some simulation experiments demonstrate the validity of our estimator.
    Download PDF (1038K)
  • Article type: Appendix
    1998 Volume 41 Issue 1 Pages 166-169
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    Download PDF (352K)
  • Article type: Appendix
    1998 Volume 41 Issue 1 Pages App2-
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    Download PDF (77K)
  • Article type: Cover
    1998 Volume 41 Issue 1 Pages Cover2-
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    Download PDF (55K)
  • Article type: Cover
    1998 Volume 41 Issue 1 Pages Cover3-
    Published: 1998
    Released on J-STAGE: June 27, 2017
    JOURNAL FREE ACCESS
    Download PDF (55K)
feedback
Top