Bandwidth packing: a tabu search approach

Article Abstract:

A tabu search method is proposed for the solution of the bandwidth packing (BWP) problem. The BWP problem occurs in telecommunications and involves call assignment to paths in a capacitated graph, while ensuring non-violation of capacities and profit maximization. The proposed tabu search approach balances intensification and diversification by incorporating search mechanisms that inhibit local optimality. The approach accomplishes this by using short-term memory-resident flexible functions that disallow inter-solution transactions that reinstate features of past solutions.This method involves a k-shortest path algorithm for obtaining optimal or near optimal solutions. Application to a 0-1 integer programming version of the BWP problem illustrates the efficiency of the method relative to others.

Telecommunications systems, Telecommunication systems

---

Tabu search and ejection chains - application to a node weighted version of the cardinality-constrained TSP

Article Abstract:

A tabu search method based on ejection chain procedures is used to solve the cardinality-constrained Traveling Salesman Problem (TSP), a special TSP problem that determines a node simple path that maximizes the sum of weights of nodes passed by, subject to lower and upper bounds on the inclusive nodes. Short term tabu search is found to improve the power of the ejection chain, while random restarting should be used conditionally rather than unconditionally to obtain optimal results. The method has also been used to seek a profitable column in a column generation procedure for solving fractional covering problems. Computational experiments confirm the importance of advanced ejection chain rules for deriving robust solutions to complicated problems.

Research, Search theory, Traveling-salesman problem

---

Adaptive memory tabu search for binary quadratic programs

Article Abstract:

A study was conducted to characterize the development and use of an adaptive memory tabu search approach in determining binary quadratic programs. The memory approach supports control mechanisms of tabu restrictions and aspiration classifications. It also considers related penalties and inducements to changing move computations. Last solutions that were produced at critical events were taken into account to guide the search by recency information. Long-term frequency data were also obtained by considering a TABU_F vector for use in the search process. Results showed that the method can be utilized to solve difficult quadratic problems. Findings also showed that the approach is better than other heuristic methods in generating solutions.

Research and Development in the Social Sciences and Humanities, Information Systems & Theory, Management Science, Methods, Analysis, Integer programming, Nonlinear theories, Heuristic programming, Nonlinear programming

Similar abstracts:

• Abstracts: The influence of supervisory promotion and network location on subordinate careers in a dual ladder RD&E setting
• Abstracts: Dell Computer battles its rivals with a lean machine: company's swift budget and price cuts led to '92 profit despite recession
• Abstracts: Haiti: Aristide's party dominates elections. Aristide, military open indirect talks. Haiti compromise sets Aristide return
• Abstracts: IBM's short list in search of a chief dwindles once again. IBM won't get two key aides of Gerstner's. IBM's new chairman shuns tradition as he puts together his own team
• Abstracts: Attempts to speed up modem connections are off to a slow start. With this little box, it's no big deal to make connections

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.