Chae Y. Lee

Parallel genetic algorithms for the earliness-tardiness job scheduling problem with general penalty weights

Abstract The purpose of this paper is to develop parallel genetic algorithms for a job scheduling problem on a single machine. The objective of the scheduling is to minimize the total generally weighted earliness and tardiness penalties from a common due date. A binary representation scheme is employed for coding job schedules into chromosomes. Parallel subpopulations are constructed by considering only jobs that can be processed first in the schedule. Three important genetic algorithm operators; reproduction, crossover and mutation are implemented by reflecting the problem-specific properties. The efficiency of the parallel genetic algorithm is illustrated with computational results.

Modeling and analysis of the dynamic location registration and paging in microcellular systems

A 0-1 integer programming model is considered to determine the most appropriate dynamic location registration (LR) area of each subscriber in microcellular systems. The minimization model of the LR area updating and paging signal costs is examined. The model is based on the subscriber characteristics, such as the call arrival rate and the velocity, as well as the regional information. The control channel blocking probabilities are considered as constraints to meet the service level of subscribers. A dynamic scheme which adaptively updates the size and shape of the LR area is developed by solving the minimization problem. Paging and location updating procedures are presented based on the dynamic procedure. The superiority of the proposed scheme is demonstrated with various computational results.

A tabu search for the survivable fiber optic communication network design

Abstract A tabu search procedure is developed to solve fiber optic communication network design problems with survivability constraints. Two systematic improving heuristics: delete-add and delete-link procedures are presented. The conditions for the candidate links to be added and deleted in the two procedures are examined by considering the feasible structures of the survivable network. A local improvement procedure is considered by combining the two heuristics for the downhill move in the search procedure. Computational results show that the proposed tabu search outperforms the best known heuristic procedure in the literature.

A design of the minimum cost ring-chain network with dual-homing survivability: a tabu search approach

Abstract This article discusses a design of the ring-chain architecture with dual homing survivability for metropolitan telecommunication networks. A self-healing ring (SHR) and multiple chains are considered to cover hub, ring nodes and other offices. Offices in a chain are connected to the ring in dual homing fashion to increase the survivability. Given a ring topology, the problem is to minimize the link cost of the chain network which satisfies the dual homing constraint. An integer programming formulation and the NP-completeness of the problem is presented. As a solution procedure, a tabu search is proposed with two types of moves; insert and swap. To increase the efficiency of the search procedure, tabulists, aspiration criteria, and diversification strategy are discussed. The computational results show that the proposed tabu search provides near optimal solutions within a few seconds. Approximately 1%–4% gap from the optimum is experienced in problems with reasonable size of metropolitan area networks.

Dynamic sectorization of microcells for balanced traffic in CDMA: genetic algorithms approach

With the increase of cellular users, traffic hot spots and unbalanced call distributions are common in wireless networks. As a solution to this problem, code-division multiple-access techniques enable a base transceiver station to connect microcells with optical fibers and to control the channels by sectorizing the microcells. To solve the load balancing among microcells, we dynamically sectorize the microcells depending on the time-varying traffic. The microcell sectorization problem is formulated as an integer linear programming that minimizes the blocked and handoff calls in the network. In the proposed sectorization, proper, connected, and compact sectors are considered to keep the handoffs as small as possible while satisfying the channel capacity at each sector. Three genetic algorithms (GAs) are proposed to solve the problem: standard GA, grouping GA, and parallel GA. Computational results show that the proposed GAs are highly effective. All three GAs illustrate outstanding performance for small size problems. The parallel GA, which is based on the operators used in grouping GA, demonstrates excellent solution quality in a reasonable time.

Balancing loads on SONET rings with integer demand splitting

SONET Self Healing ring is one of the most intriguing schemes which provide survivability for telecommunication networks. To design a cost effective SONET ring it is necessary to consider load balancing problems by which the link capacity is determined. In this paper, we consider the load balancing problem in SONET ring when demand splitting is allowed only by integer. A bounded approximation to the optimal solution is presented and proved by employing the arc-demand matrix which represents the amount of demand routed on each arc. A very efficient algorithm is provided which always satisfies the optimality at termination. The algorithm iteratively decreases the maximum load by considering counter-clockwise routing of a demand which passes through the arc with maximum load. Justification of the algorithm is also provided.

A lexicographically fair allocation of discrete bandwidth for multirate multicast traffics

Fair bandwidth allocation is an important issue in the multicast network to serve each multicast traffic at a fair rate commensurate with the receivers capabilities and the capacity of the path of the traffic. Lexicographically fair bandwidth layer allocation problem is considered and formulated as a nonlinear integer programming problem. A nonincreasing convex function of the bandwidth layers of the virtual sessions is employed to maximize the bandwidth of each virtual session from the smallest. To solve the fairness problem a genetic algorithm (GA) is developed based on the fitness function, ranking selection and the shift crossover. Outstanding performance is obtained by the proposed GA in various multicast networks. The effectiveness of the GA becomes more powerful as the network size increases.

Discrete bandwidth allocation considering fairness and transmission load in multicast networks

Abstract As a promising solution to tackle the network heterogeneity in multicasting, layered multicast protocols such as receiver-driven layered multicast (RLM) and layered video multicast with retransmission (LVMR) have been proposed. This paper considers fairness as well as transmission load in the layered multicasting. Lexicographically fair bandwidth allocation among multicast receivers is considered under the constraint of minimum bandwidth requirement and the link capacity of the network. The problem of transmission load in the layer multicasting due to various user requirements is also examined by minimizing the number of layers. The bandwidth allocation is formulated as a nonlinear integer programming problem. A dual-objective tabu search is proposed to solve the fairness and transmission load problem in multicast networks. Outstanding performance is obtained by the proposed tabu search. When the fairness objective is considered, the solution gap from the optimal solution is less than 2% in problems with 50 virtual sessions. The complexity of the dual objective largely depends on the weighting factor of the two objectives. Even in tough cases the proposed tabu search provides excellent solution, whose gap is within 6% from the optimal solution.

Effective Video Multicast Using SVC with Heterogeneous User Demands over TDMA-Based Wireless Mesh Networks

We provide an effective video multicast framework over time division multiple access (TDMA)-based wireless mesh networks (WMNs). Heterogeneous user demand is considered where each multicast receiver has its own video quality demand. In video multicasting, different relay nodes in the same multicast tree may have different transmission rates by utilizing scalable video coding (SVC). Considering this distinguishing characteristic of video multicasting, the proposed multicast framework provides effective multicast routing, scheduling, and rate allocation algorithms. The purpose of the multicast routing and scheduling is to obtain a minimum length schedule which satisfies given quality demands of receivers. If the schedule is not feasible even with its minimum length due to the limited number of time slots in the network, rate allocation algorithm adjusts the transmission rates of relay nodes to generate a feasible schedule. The algorithm is designed to maximize the minimum utility of multicast receivers. Simulation results show that the proposed multicast routing and scheduling algorithms outperform existing schemes in terms of schedule length. The minimum utility is improved up to 30 percent by the proposed rate allocation algorithm compared to the existing method.

A parametric power control with fast convergence in cellular radio systems

The capacity of a cellular radio system is largely dependent on its transmitter power control. Since power control is inherently a real-time problem, to find the fastest carrier-to-interference ratio (CIR) balancing algorithm, which forces the CIR of each cell to converge to a value, has been the essential issue. An efficient parametric power control (PPC) scheme is developed in this paper. In this scheme, the power control is performed at each base by using some parameters provided by the central collector, which determines the multiplier of the power update function. The algebraic property of its CIR balancing algorithm is analyzed. In an environment with zero noise, the scheme proposes a quick method for obtaining a least upper bound on the achievable CIR. The proposed scheme PPC is also considered in a cellular system with positive receiver noise. The computational results show that the convergence of the proposed CIR balancing algorithm is quick and the power consumption is reasonable compared to distributed schemes. With the proposed algorithm, the CIRs are balanced sufficiently in a short power control period.