Woo-Tsong Lin

Ant colony optimization-based algorithm for airline crew scheduling problem

Research highlights? We formulate airline crew scheduling problem as Traveling salesman problem with constrained and then introduce Ant Colony Optimization algorithm to solve it. ? Performance of the proposed ACO-based algorithm is examined on real cases of airline companies. ? Ant Colony Optimization algorithm (ACO) perform more effective and robust than Genetic algorithms for airline crew scheduling problem. Airline crew scheduling is an NP-hard constrained combinatorial optimization problem, and an effective crew scheduling system is essential for reducing operating costs in the airline industry. Ant colony optimization algorithm (ACO) has successfully applied to solve many difficult and classical optimization problems especially on traveling salesman problems (TSP). Therefore, this paper formulated airline crew scheduling problem as Traveling Salesman Problem and then introduce ant colony optimization algorithm to solve it. Performance was evaluated by performing computational tests regarding real cases as the test problems. The results showed that ACO-based algorithm can be potential technique for airline crew scheduling.

Markowitz-based portfolio selection with cardinality constraints using improved particle swarm optimization

This work presents particle swarm optimization (PSO), a collaborative population-based meta-heuristic algorithm for solving the Cardinality Constraints Markowitz Portfolio Optimization problem (CCMPO problem). To our knowledge, an efficient algorithmic solution for this nonlinear mixed quadratic programming problem has not been proposed until now. Using heuristic algorithms in this case is imperative. To solve the CCMPO problem, the proposed improved PSO increases exploration in the initial search steps and improves convergence speed in the final search steps. Numerical solutions are obtained for five analyses of weekly price data for the following indices for the period March, 1992 to September, 1997: Hang Seng 31 in Hong Kong, DAX 100 in Germany, FTSE 100 in UK, S&P 100 in USA and Nikkei 225 in Japan. The test results indicate that the proposed PSO is much more robust and effective than existing PSO algorithms, especially for low-risk investment portfolios. In most cases, the PSO outperformed genetic algorithm (GA), simulated annealing (SA), and tabu search (TS).

Citation analysis and bibliometric approach for ant colony optimization from 1996 to 2010

To build awareness of the development of ant colony optimization (ACO), this study clarifies the citation and bibliometric analysis of research publications of ACO during 1996-2010. This study analysed 12,960 citations from a total of 1372 articles dealing with ACO published in 517 journals based on the databases of SCIE, SSCI and AH&CI, retrieved via the Web of Science. Bradford Law and Lotkas Law, respectively, examined the distribution of journal articles and author productivity. Furthermore, this study determines the citation impact of ACO using parameters such as extent of citation received in terms of number of citations per study, distribution of citations over time, distribution of citations among domains, citation of authors, citation of institutions, highly cited papers and citing journals and impact factor of 12,960 citations. This study can help researchers to better understand the history, current status and trends of ACO in the advanced study of it.

Ant Colony Optimization for Markowitz Mean-Variance Portfolio Model

This work presents Ant Colony Optimization (ACO), which was initially developed to be a meta-heuristic for combinatorial optimization, for solving the cardinality constraints Markowitz mean-variance portfolio model (nonlinear mixed quadratic programming problem). To our knowledge, an efficient algorithmic solution for this problem has not been proposed until now. Using heuristic algorithms in this case is imperative. Numerical solutions are obtained for five analyses of weekly price data for the following indices for the period March, 1992 to September, 1997: Hang Seng 31 in Hong Kong, DAX 100 in Germany, FTSE 100 in UK, S&P 100 in USA and Nikkei 225 in Japan. The test results indicate that the ACO is much more robust and effective than Particle swarm optimization (PSO), especially for low-risk investment portfolios.

An Object-Oriented Architecture for Knowledge-based Production Scheduling Systems

This paper describes an object-oriented architecture to support decision making in production scheduling environments. An object-oriented world view is used to integrate concepts from discrete event simulation, conventional scheduling logic and artificial intelligence to produce capacity-feasible schedules. The architecture was implemented as a collection of loosely coupled reusable software objects by extending the functionality of software objects from BLOCS/M (Berkeley Library of Objects for Control and Simulation of Manufacturing). Our experience with an industrial prototype is presented.

A DECISION MODEL FOR REVERSE LOGISTICS SERVICE PROVIDERS IN DETERMINING ROBUST OPTIMAL PROCESSING QUANTITIES OF RETURNED PRODUCTS

ABSTRACT Reverse logistics covers a serial of activities in dealing with returned products from consumers, including collecting, reusing and recycling. Implementing reverse logistics is much more complicated and expensive than forward logistics to an enterprise. Meanwhile, the systematic patterns for handling transportation, storage, processing and management processes of these activities are still called for. Consequently, to reduce the reverse logistics cost and focus on its core business, an enterprise prefers outsourcing these activities in this manner. Previous studies focused on the selection of processing facilities and the infrastructure design of reverse logistics distribution channels for third-party reverse logistics service providers. In contrast, this research aims to deal with the issues of reverse logistics from different viewpoint. We propose a decision model for a reverse logistics service provider under the context of uncertain, multi-period, multi-type returned/recycled products and multiple processing facilities environment. The major focus of this model is on determining the robust optimal quantities of customer orders and robust optimal processing quantities of returned products for each processing facility. To deal with the issues of uncertainties, the model applies the scenario-based robust optimization approach. Further information on experiment results and implications can be found in this paper.

A Two-Dimensional Genetic Algorithm and Its Application to Aircraft Scheduling Problem

Genetic algorithms have become increasingly important for researchers in resolving difficult problems because they can provide feasible solutions in limited time. Using genetic algorithms to solve a problem involves first defining a representation that describes the problem states. Most previous studies have adopted one-dimensional representation. Some real problems are, however, naturally suitable to two-dimensional representation. Therefore, a two-dimensional encoding representation is designed and the traditional genetic algorithm is modified to fit the representation. Particularly, appropriate two-dimensional crossover and mutation operations are proposed to generate candidate chromosomes in the next generations. A two-dimensional repairing mechanism is also developed to adjust infeasible chromosomes to feasible ones. Finally, the proposed approach is used to solve the scheduling problem of assigning aircrafts to a time table in an airline company for demonstrating the effectiveness of the proposed genetic algorithm.

Swarm intelligence for cardinality-constrained portfolio problems

This work presents Particle Swarm Optimization (PSO), a collaborative population-based swarm intelligent algorithm for solving the cardinality constraints portfolio optimization problem (CCPO problem). To solve the CCPO problem, the proposed improved PSO increases exploration in the initial search steps and improves convergence speed in the final search steps. Numerical solutions are obtained for five analyses of weekly price data for the following indices for the period March, 1992 to September, 1997: Hang Seng 31 in Hong Kong, DAX 100 in Germany, FTSE 100 in UK, S&P 100 in USA and Nikkei 225 in Japan. The computational test results indicate that the proposed PSO outperformed basic PSO algorithm, genetic algorithm (GA), simulated annealing (SA), and tabu search (TS) in most cases.

THE RESEARCH ON THE UNCERTAINTY FACTORS AND THE ADAPTIVE STRATEGIES FOR THE IMPLEMENTATION OF SUPPLY CHAIN MANAGEMENT—USING INFORMATION AND ELECTRONIC INDUSTRY IN TAIWAN AS THE STUDY OBJECT

ABSTRACT There are many critical successful factors for implementing supply chain management, and the ability to manage the impact of uncertainty in the supply chain management is one of them. This study was an exploratory research, which tried to categorize the sources and impacts of uncertainty factors based on the literature review in the first place. Then, a prototype of the architecture of adaptive strategies for uncertainty factors was proposed, which used IDEF0, a graphics modeling tools, as a mean of representing the business processes for finding out the sources and the impacts that a company suffered by uncertainty factors. Further, this study took case study as the research method to: (1) understand how the information and electronic companies in Taiwan did as they suffered from uncertainty; and (2) modify the prototype of the proposed architecture for better fitting the real need of the industry. Finally, this study synthesized the discoveries, and provided helpful suggestions to the information and electronic industry in Taiwan for a reference if they want to implement supply chain management.

Innovative chance discovery for finding innovative useful idea in travelling

If the following enterprises want to enter the blue ocean strategy market with high profits, they need to find the value for new products earlier and take up a position in the new markets. In order to survive for a longer time, enterprises have to find the rare and valuable information at the early stage. To achieve this purpose, the researchers proposed a BCG-based novel innovation classification model. We use KeyGraph algorithms to find the rare keyvalue and the inverse cluster frequency algorithm proposed by this study to conduct two-objective decision analysis. Then, we use the Boston consulting group model to classify keyvalue and inverse cluster frequencys value in which we get the early innovative value. Finally, we used the innovative value to test college freshmen to see the difference between before and after the college freshmen accepted the innovative value. The results show that 59% of the college freshmen perceive the usefulness of the new product, which demonstrates a significant progress. This is to justify that the model proposed by this study can be of help to mine the rare and valuable information early and effectively.