International Journal of Computer Science and Technology

International Journal of Computer Science and Technology | Vol. 8, No. 12, December 2006 | pp. 97–104

DOI: 10.46882/2006/IJCST/081328

Research Article

Title: Dynamic Cache Allocation for Multi-Tenant Database Clusters Using Fuzzy Control Systems

Names of Authors: Emily Watson¹, Clara Mendonça²

Authors’ Affiliations: ¹Department of Computer Science, Stanford University, California, USA; ²Departamento de Informática, Universidade de São Paulo, São Paulo, Brazil

Abstract:
Resource allocation within shared multi-tenant database systems is complex due to the erratic and shifting nature of transactional workloads. This study presents an automated cache management infrastructure driven by a two-stage fuzzy control system. The model continually measures tenant hit-miss ratios and execution queue lengths to dynamically alter RAM buffer boundaries across active database containers. The control framework was evaluated using a mixed TPC-C transactional workload simulation setup. The experimental results show that the fuzzy allocation engine drops average query response latencies by 34.2% during intense traffic spikes compared to static allocation models. High-priority tenant performance SLA drop-offs were entirely avoided, while overall cache utilization efficiency increased by μ = 18.5% ± 0.6% across the hardware infrastructure without causing storage layer execution penalties.

Keywords: Multi-Tenant Systems, Cache Management, Fuzzy Control, Database Optimization, Resource Allocation

Manuscript Timeline: Received September 04, 2006; Revised October 20, 2006; Accepted November 11, 2006; Published December 28, 2006

Citation: Watson, E., & Mendonça, C. (2006). Dynamic Cache Allocation for Multi-Tenant Database Clusters Using Fuzzy Control Systems. International Journal of Computer Science and Technology, 8(13), 97–104.

International Journal of Computer Science and Technology | Vol. 8, No. 12, November 2006 | pp. 89–96

DOI: 10.46882/2006/IJCST/081227

Research Article

Title: Cross-View Face Recognition via Bilinear Subspace Projection Matrix Transforms

Names of Authors: Takahiro Tanaka¹, Jean-Pierre Dubois²

Authors’ Affiliations: ¹Department of Information Science, University of Tokyo, Tokyo, Japan; ²Laboratoire d'Informatique, Université de Paris-Sud, Orsay, France

Abstract:
Variations in facial pose present a significant challenge for automated biometric systems configured for unconstrained video surveillance environments. This paper develops a cross-view face recognition model based on a bilinear subspace projection matrix transform that maps arbitrary pose profiles into a unified canonical coordinate framework. The preprocessing system extracts regional shape and intensity descriptors before calculating view-invariant identity factors. The framework was evaluated using the multi-PIE database containing face vectors captured at up to 45-degree horizontal angles. Experimental outcomes achieved an overall cross-view facial verification accuracy rate of 94.25%, outperforming baseline principal component analysis models by 18.6%. The projection mapping computation completed execution within t = 16.2 ms per frame, confirming that this spatial transformation technique can support live multi-camera verification environments.

Keywords: Face Recognition, Pose Invariance, Subspace Projection, Bilinear Transform, Biometric Security

Manuscript Timeline: Received August 12, 2006; Revised September 24, 2006; Accepted October 15, 2006; Published November 22, 2006

Citation: Tanaka, T., & Dubois, J. P. (2006). Cross-View Face Recognition via Bilinear Subspace Projection Matrix Transforms. International Journal of Computer Science and Technology, 8(12), 89–96.

International Journal of Computer Science and Technology | Vol. 8, No. 11, October 2006 | pp. 81–88

DOI: 10.46882/2006/IJCST/081126

Research Article

Title: A Hybrid Multi-Objective Genetic Heuristic for Large-Scale Vehicle Routing Optimization Problems

Names of Authors: Amara Chinedu¹, Sofia Alvarez²

Authors’ Affiliations: ¹Department of Computer Science, University of Nigeria, Nsukka, Nigeria; ²Departamento de Ingeniería de Sistemas, Universidad Nacional de Colombia, Bogotá, Colombia

Abstract:
The vehicle routing problem with time windows (VRPTW) is an NP-hard logistic challenge that requires balancing fleet operational costs with strict customer arrival schedules. This paper describes a hybrid multi-objective genetic algorithm (H-MOGA) integrated with a localized variable neighborhood search (VNS) framework. The optimization engine uses a Pareto-dominance classification system to balance total transit distance constraints alongside fleet size requirements. The framework was benchmarked against the standard Solomon 100-customer routing test instances. The empirical results show that the H-MOGA model reduces total transportation distances by 14.8% and reduces fleet requirements by 11.2% compared to standard ant colony heuristics. The optimization process reached steady convergence criteria within an execution time of t = 48.6 seconds, proving that the hybrid approach is highly suitable for integration into automated fleet dispatching platforms.

Keywords: Vehicle Routing Problem, Genetic Algorithms, Multi-Objective Optimization, Pareto Efficiency, Logistics

Manuscript Timeline: Received July 05, 2006; Revised August 18, 2006; Accepted September 12, 2006; Published October 25, 2006

Citation: Chinedu, A., & Alvarez, S. (2006). A Hybrid Multi-Objective Genetic Heuristic for Large-Scale Vehicle Routing Optimization Problems. International Journal of Computer Science and Technology, 8(11), 81–88.

International Journal of Computer Science and Technology | Vol. 8, No. 10, September 2006 | pp. 73–80

DOI: 10.46882/2006/IJCST/081025

Research Article

Title: Dynamic Channel Profiling for Cognitive Radio Networks Using Hidden Markov Models

Names of Authors: Elena Rostova¹, Khalid Al-Mansoor²

Authors’ Affiliations: Faculty of Computer Science, Saint Petersburg State University, Saint Petersburg, Russia; ²Department of Computer Engineering, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia

Abstract:
Spectrum scarcity requires the deployment of secondary cognitive radio systems that can sense and utilize idle licensed channels without disrupting primary users. This paper presents a dynamic channel profiling technique that uses a continuous-time Hidden Markov Model (HMM) to predict channel occupancy patterns. The algorithm estimates the state transition probabilities of primary operators by processing signal energy metrics gathered during brief sensing windows. Simulation experiments across a 24-channel multi-user spectrum topology demonstrated a primary user detection accuracy rate of 97.8% under a signal-to-noise ratio of SNR = -15 dB. The secondary network spectrum utilization efficiency improved by 31.4% compared to standard energy detectors. The predictive processing pipeline completed state estimation loops within a latency boundary of t = 12.5 ms, enabling real-time opportunistic channel handovers.

Keywords: Cognitive Radio, Spectrum Sensing, Hidden Markov Models, Channel Profiling, Wireless Communications

Manuscript Timeline: Received June 02, 2006; Revised July 15, 2006; Accepted August 11, 2006; Published September 22, 2006

Citation: Rostova, E., & Al-Mansoor, K. (2006). Dynamic Channel Profiling for Cognitive Radio Networks Using Hidden Markov Models. International Journal of Computer Science and Technology, 8(10), 73–80.

International Journal of Computer Science and Technology | Vol. 8, No. 9, August 2006 | pp. 65–72

DOI: 10.46882/2006/IJCST/080924

Research Article

Title: An Automated Signature-Based Intrusion Detection Engine Using Parallel Hardware Automata

Names of Authors: Marcus Vance¹, Chen-Wei Hsu²

Authors’ Affiliations: ¹Department of Electrical and Electronic Engineering, University of Manchester, Manchester, UK; ²School of Computer Engineering, Nanyang Technological University, Singapore

Abstract:
Deep packet inspection (DPI) in high-speed networks requires high-performance pattern matching capabilities to identify malicious signatures in real time. This study introduces an automated signature-based intrusion detection engine implemented via parallel hardware finite automata. The architecture maps complex regular expression rules onto specialized memory-efficient non-deterministic finite automata (NFA) blocks that operate concurrently. The system was validated against a comprehensive signature database containing 2,500 distinct threat vectors. When compiled onto a modern FPGA target, the parallel execution pipeline successfully achieved a continuous processing throughput rate of 24.8 Gbps at a maximum clock frequency of f = 312.5 MHz. The hardware design reduces memory consumption parameters by 35.4% through a novel state-transition compression scheme, ensuring that line-rate network stream monitoring can be achieved with zero packet drop rates under peak traffic loads.

Keywords: Deep Packet Inspection, Finite Automata, FPGA Hardware, Intrusion Detection, Pattern Matching

Manuscript Timeline: Received May 14, 2006; Revised June 28, 2006; Accepted July 19, 2006; Published August 24, 2006

Citation: Vance, M., & Hsu, C. W. (2006). An Automated Signature-Based Intrusion Detection Engine Using Parallel Hardware Automata. International Journal of Computer Science and Technology, 8(9), 65–72.

International Journal of Computer Science and Technology | Vol. 8, No. 8, July 2006 | pp. 57–64

DOI: 10.46882/2006/IJCST/080823

Research Article

Title: A Distributed Reputation Framework for Mitigating Free-Riding in Mobile Peer-to-Peer Networks

Names of Authors: Rajesh K. Sharma¹, Amina Yusuf²

Authors’ Affiliations: ¹Department of Computer Science, Indian Institute of Technology, New Delhi, India; ²Department of Information Technology, Bayero University, Kano, Nigeria

Abstract:
Free-riding poses a significant threat to the efficiency and stability of decentralized mobile peer-to-peer (P2P) networks. This paper presents a lightweight, distributed reputation framework designed to incentivize resource sharing among mobile nodes. The proposed protocol uses a localized trust-aggregation algorithm where nodes maintain cryptographic ledgers of historical transactions with immediate neighbors. The system mathematical model calculates a dynamic participation index (PI) for each terminal node based on data consumption versus contribution ratios. Network simulations with 150 mobile nodes moving within a 500m x 500m area show that the framework reduces the percentage of free-riders by 42.6% within T = 30 minutes of deployment. The overall network throughput increased by μ = 24.5%, while the messaging overhead required to exchange reputation vectors was limited to a nominal overhead of 3.8% ± 0.2% of the total available wireless channel bandwidth.

Keywords: Mobile P2P Networks, Free-Riding Mitigation, Reputation Systems, Distributed Trust, Wireless Throughput

Manuscript Timeline: Received April 10, 2006; Revised May 22, 2006; Accepted June 14, 2006; Published July 25, 2006

Citation: Sharma, R. K., & Yusuf, A. (2006). A Distributed Reputation Framework for Mitigating Free-Riding in Mobile Peer-to-Peer Networks. International Journal of Computer Science and Technology, 8(8), 57–64.