Applying swarm intelligence to a library system

4 downloads 17277 Views 1MB Size Report
Jan 22, 2010 - intelligence to develop friendly human-computer interface software for readers using a personal or notebook computer. We program the system ...
Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

Contents lists available at ScienceDirect

Library Collections, Acquisitions, & Technical Services j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / l c a t s

Applying swarm intelligence to a library system Li-Shan Chen ⁎ Senior Lecturer, Department of Information Management, Fortune Institute of Technology, No.1-10, Nongchang Rd, Daliao Township, Kaohsiung County 83160, Taiwan (ROC)

a r t i c l e

i n f o

Available online 22 January 2010 Keywords: Software development Integrated library systems Internet access and use User interface design

a b s t r a c t This paper aims to integrate a library system so that it becomes intelligent. We use swarm intelligence to develop friendly human-computer interface software for readers using a personal or notebook computer. We program the system and software with Extensible Markup Language (XML) and C Sharp language. The kernel library automatically communicates with other libraries by agents, so readers can search from the closest library. This study adds only one component to the kernel library, and the other libraries do not add this component. They maintain their original status. Readers do not use a browser; they directly communicate with the library search system, saving much time. Readers without IT skills can also easily search for books in the library system. © 2009 Elsevier Ltd. All rights reserved.

1. Introduction Digital inequality is one of the most critical issues in the knowledge economy. The private and public sectors have devoted tremendous resources to address such inequality, yet the results have been inconclusive (Hsieh et al., 2008) [1]. The use of digital resources by humanities scholars has provided strong evidence of the continuing importance of both physical and digital information resources (Warwick et al., 2008) [2]. The majority of respondents do not seek help from library staff, and libraries are mainly regarded as places to borrow books rather than purveyors of electronic information. Traditional and electronic library use was highest amongst higher education students (Myhill, 2007) [3]. The digitization of material within libraries and archives has become commonplace as technology can now offer better quality image reproduction. The primary reasons for an institution to start a digitization project are to preserve material and widen access. However, both the potential and the risks remain a matter of debate (Petrelli and Auld, 2008) [4]. Mansourian (2008a and 2008b) [5,6] reported on a wider study of the interaction of end users with search tools on the World Wide Web (hereafter the Web). The initial focus of the main study was on the invisible Web, i.e., parts of the Web that general purpose search engines either could not or were not intended to index. However, as it was exploratory research, some new directions emerged that have been addressed in a series of publications. One of these emergent directions is addressed in this paper, to explore users' reactions to failure in searching the Web. Lazarinis (2007) [7] provided learners with the required knowledge to access a search engine, to formulate their queries, to quickly evaluate the results, and to navigate in the retrieved set of documents. Nowadays, every library system is independent, and the traditional searching method suffers disadvantages in actual usage. For example, different units have different book categories and book volumes. When readers do not find what they need at a unit's Web site, they have to go to different units for further searches until they obtain what they need. Therefore, the above process is a series of repeated search actions which is inefficient and time-consuming. If readers do not use a browser, they can directly communicate with the library search system, saving much time. Readers without IT skills can easily search for books in the library system. To solve the above problems, this study uses an ant algorithm to

⁎ Fax: +886 7788 9777. E-mail addresses: [email protected], [email protected]. 1464-9055/$ – see front matter © 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.lcats.2009.11.002

2

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

develop an intelligent library searching system. The system can communicate with other library systems synchronously. These library systems become allied. This study also develops human–computer interface software embedded in readers' machines (e.g., personal or notebook computers). The software's size is only 64 kilobits, which is small. This study lets the searching process become intelligent. As long as readers can get online, they can easily and conveniently obtain related book information from the library that is closest to them. This will be very helpful in obtaining library resource information. 2. Literature review 2.1. Library and information science Chen and Chen (2007) [8] successfully developed a knowledge base of intelligent systems. Duan et al. (2009) [9] empirically examined informational cascades in the context of online software adoption, and found users' behavior in adopting software products was consistent with the predictions of the informational cascades literature. Abbasi and Chen (2008) [10] proposed a design framework for computer-mediated communication text analysis systems. Feller et al. (2008) [11] highlighted the interplay between digital and social networks within open source software networks, demonstrating that the use of social mechanisms was inherently dependent upon the underlying IT infrastructure. Agarwal et al. (2008) [12] shed light on the technical, behavioral, and economic challenges and implications of such networks, contributing to our understanding of how their power can be harnessed. Bapna et al. (2009) [13] developed three auction-based pricing and allocation solution methods for a case where a capacityconstrained online service provider offered multiple classes of unique, one-time services with differentiated quality. Hussain and Cornelius (2009) [14] highlighted how an appropriate, sophisticated use of what Giddens referred to as the duality of structure contributed to the consolidation of an IT manager's credibility and authority. Larsen et al. (2009) [15] used a socio-technical approach to consider UML as a technology embedded in a social environment, and the project developers were interviewed in detail about their use of UML along with influences on their decisions to use this approach and the results of using it. Gilstrap (2009) [16] proposed a complex systems framework for future research on leadership and organizational development surrounding change in academic libraries and professional responsibilities. Saunders (2009) [17] surveyed 13 information literacy experts about proposed futures to explore the possible evolution of information literacy over the next decade. Saunders (2008) [18] proved that access mechanisms were very important predictors of information resource satisfaction, but library facilities and library staff were negligible predictors. Dana and Sandra (2009) [19] explored the perceived information and communication technology (ICT) competencies of students enrolled in school library certification programs and the use of ICTs in their school library education programs. Aharony (2009) [20] explored whether librarians, whose main work focuses on information, were familiar with new technological changes and innovations, and whether they made use of Web 2.0 applications. Chung (2009) [21] presented an analysis model for setting up a core journal collection for academic libraries. Pomerantz and White (2009) [22] recognized a need to alter acquisitions procedures and codes for more effective use of the Innovative Interfaces Millennium acquisitions module for budget management. Chen (2008) [23] combined swarm intelligence and Web services to transform a conventional library system into an intelligent library system with high integrity, usability, correctness, and reliability software for readers. Duinkerken et al. (2008) [24] described the Texas A&M University (TAMU) Library's new and innovative approach to monograph collection development using a reengineered internal funding structure and a process that reflects the needs and goals of the library and its users. Adkins et al. (2008) [25] suggested that library and information science education, beyond supporting leisure reading, ought to emphasize the roles that leisure reading serves for readers. Termens (2008) [26] showed that some universities make more use of them than others, taking into account their relative potential in terms of full-time equivalent (FTE) faculty members, and the methodological model proved to be viable for studying patterns of use at more detailed levels than the general institutional level normally covered by COUNTER-compliant reports. Quinn (2008) [27] examined psychological research on group decision making and explored how group psychology influences decision making and what the implications might be for collection development. Martinovic and Cukic (2008) [28] performed faster character recognition from printed media and their forwarding to a Web library, and enabled the setting of execution conditions, mapping of pictures with characters to clients, and execution monitoring. 2.2. Swarm intelligence Recently, nature-inspired intelligence techniques have become attractive for analyzing large data sets and solving complex optimization problems. Ant colony optimization (ACO) can be applied to the data mining field to extract rule-based classifiers (Martens et al., 2007) [29]. It is a promising metaheuristic technique, and a great amount of research has been devoted to its empirical and theoretical analyses (Birattari et al., 2007) [30]. The learners and lecturers agreed that style-based ant colony systems could provide useful supplementary learning paths (Wang et al., 2008) [31]. A novel hybrid optimization algorithm based on ant colony optimization has been developed and applied to target motion analysis, pap-smear cell classification, and tool path problems. Its effectiveness is comparable to those of standard estimators (Nolle, 2008 [32], Marinakis and Dounias, 2008 [33], and Tewolde and Sheng, 2008 [34]).

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

3

3. Methodology The intelligent library searching system was developed in the environment of: Microsoft Windows Server 2008, Internet Information Services 7.0 (IIS 7.0), Microsoft Structured Query Language (MS-SQL) Server 2008, and Visual Studio 2008 (VS2008). The human–computer interface software was developed in the environment of Microsoft Vista Ultimate, MS-SQL Server 2008, and edited on VS 2008. The programming language is Extensible Markup Language(XML) and C Sharp language. 3.1. Structure This study does not change the original structure of the library system. It adds only one component, Library Querying System Modeling Base, to the library system, and sets the kernel library at the Fortune Institute of Technology (F.I.T.) library. The searching path is from (City)i to (City)j, and the distance of (City)i to (City)j is the closest, as shown in Fig. 1. The searching path begins at (City)1, and it has two choices. One is (City)2, and the other is (City)3. Because the distance from (City)1 to (City)2 is shorter than the distance from (City)1 to (City)3, the optimum searching path is (City)1 to (City)2. Likewise, the optimal searching path is “(City)1 →(City)2 →(City)3 →(City)4 →(City)5 →(City)6 →…” The search process is shown in Fig. 2. The framework of the library system is shown in Fig. 3. The readers are divided into two types. One is the school staff and students, and the other is citizens. 3.2. Principle This study amends the ant algorithm of Birattari et al. (2007) [30] so that the library system becomes intelligent and mobile. The design of the “Library Querying System Modeling Base” is very important, and it is the kernel technology in this study. The development process is described below. (Note: The agents are seen as the ants.) 3.2.1. Meaning of the symbols and nouns (a) n: Then numbers of cities. P (b) m = bi ðt Þ: The total agents. i=1

bi(t): The numbers of agents in the (City)i. (c) dij: The distant between (City)i and(City)j. This study considers that it is symmetrical; therefore, dij is equal to dji. (d) τij(t): The intensity of the pheromone upper edge. τij ðt Þ = ρ τ ij ðt Þ + Δτ ij

ð1Þ

This study uses (Eq. 1) to update the pheromone. ρ: The parameters of pheromone evaporation ðeÞ Δτij =

m X

k

Δτ ij

ð2Þ

k=1

Δτkij: The kth agent remains a pheromone going through the edge (i, j). It is defined as equation 3. Q: The influential parameter of the pheromone.

Fig. 1. Searching path.

4

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

Fig. 2. Searching process of readers.

Lk: The total length of the route; the kth agent goes all over the city. Δτ kij = LQK , The kth agent goes through edge (i, j) between time points t and (t + ti). k

ð3Þ

Δτij = 0; otherwise

Fig. 3. Framework of intelligent querying system.

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

5

(f) R: The cycle's counter agent goes through all the cities, and Rmax is the upper limit of R. (g) Tabuk(I): The record of the kth that has gone through the cities. The “I” is to make a visit to “Ith” city. This can prevent the agent from going back to cities already visited. (h) μij: The inverse of the distance between (City)i and (City)j. μ ij =

1 dij

ð4Þ

(i) pkij(t): The probability that the kth agent goes from (City)i to (City)j.

k Pij ðt Þ

8 > > > > > > > > > < =

h

iα h iβ τij ðt Þ μ ij ½τ ðtÞα ½μ β ; if jaðn − Tabukðt ÞÞ

 P > ik ik > > > ka n− n − TabukðIÞ > > > > > : 0; otherwise

ð5Þ

α and β are the important controlling parameters of pheromone information and μij.

3.2.2. “Library Querying System Modeling Base” designs process The “Library Querying System Modeling Base” is the intelligent agents' generator. It is the core of the library system. The kernel library communicates with other library systems through it. The design steps are described below. Step 1: Set t = 0, R = 0 (“t” is the time counter, and “R” is the cycles counter). For all edge (i, j), set τij(t) = constant, Δτij = 0. To put m agents into n cities, Step 2: Set I = 1 (”I” is Tabu list index). For k = 1 to m (The record of the kth agent is listed in Tabuk (I) at initial city.), Step 3: Set I = I +1. For k = 1 to m (Use Equation A.5 to decide (City)j and moving the kth agent to (City)j recorded in Tabuk(I).), Step 4: For k = 1 to m do To move the kth agent from Tabuk (n) to Tabuk (1) and calculate the total length of all paths recorded, and update the shortest path.

Fig. 4. Entrance.

6

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

Fig. 5. Operating frame.

Fig. 6. Searching results in “step 2.”

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

Fig. 7. Detail of “ATM & MPLS th…” in step 3.”

Fig. 8. Operating frame.

7

8

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

Fig. 9. Searching results in “step 2.”

Fig. 10. Detail of “CCIE study guide” in “step 3.”

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

9

To calculate each edge (i, j), for k = 1 to m do

k

Δτij =

f

Q Lk

if ði; jÞaTabuk 0;

otherwise k

Δτ ij = Δτij + Δτij Step 5: By τij(t + t1) = ρτij(t) + Δτij, calculates τij(t + t1) for each edge (i, j). Set t = t1, R = R + 1. Set Δτij = 0 for each edge (i, j). Step 6: If (R b Rmax) and (no entering in stop situation), then clear the entire Tabu list. Go to step 2. Else print the shortest path and stop. 4. Application in libraries Fig. 4 is the entrance frame. There are two buttons. One is for students and school staff, and the other is for citizens. If the readers are students, or school staff, they will click the upper button to enter the Fig. 5 frame. They can key in the keyword in “Step 1” (for example, key in “OS”). Then the agents will search for books in “FIT Library” from the closest library, and the search results will be shown in “Step 2,” as shown in Fig. 6. If the readers want to understand the details of the results (for example, the title is “ATM & MPLS...”), they can click the row result “ATM & MPLS...” to view its detail in “Step 3,” as shown in Fig. 7. If the readers are citizens, they will click the lower button in Fig. 4 to enter the Fig. 8 frame. There are two items in “Step 1.” One is the citizen's location, and the other is the “keyword of title” which the citizen wants to search. In this case, the citizen is at Kaohsiung and he (or she) wants to search the “CCIE.” The search results are shown in “Step 2,” as shown in Fig. 9. If the citizen wants to understand the details of the results (for example, the title is “CCIE study guide”), they can click the row result “CCIE study guide” to view its detail in “Step 3,” as shown in Fig. 10. 5. Conclusion This study adopts and programs swarm intelligence into a library system, so the system can become intelligent. The kernel library automatically communicates with other libraries by agents, so readers can search from the closest library using an ant colony system. Readers can be in any city. This study only adds one component to the kernel library, and the other libraries do not add this component. The other libraries maintain their original status. As long as the readers can get online, they can easily and conveniently obtain book information in the library that is closest to them. This will be very helpful in obtaining library resource information. The human–computer interface software is designed to cross platforms. Its size is only 64 kilobits, so it is not a burden to readers' equipment. The software is high in integrity, usability, correctness, and reliability. References [1] Hsieh, J. J., Rai, A., & Keil, M. (2008). Understanding digital inequality: comparing continued use behavioral models of the socio-economically advantaged and disadvantaged. MIS Quarterly, 32 (1), 97–126. [2] Warwick, C., Terras, M., Galina, I., Huntington, P., & Pappa, N. (2008). Library and information resources and users of digital resources in the humanities. Program: Electronic Library and Information Systems, 42 (1), 5–27. [3] Myhill, M. (2007). Canute rules the waves? Hope for e-library tools facing the challenge of the “Google generation. Program: Electronic Library and Information Systems, 41 (1), 5–19. [4] Petrelli, D., & Auld, D. (2008). An examination of automatic video retrieval technology on access to the contents of an historical video archive. Program: Electronic Library and Information Systems, 42 (2), 115–136. [5] Mansourian, Y. (2008). Coping strategies in Web searching. Program: Electronic Library and Information Systems, 42 (1), 28–39. [6] Mansourian, Y. (2008). Web search efficacy: definition and implementation. Aslib Proceedings, 60 (4), 349–363. [7] Lazarinis, F. (2007). Forming an instructional approach to teach Web searching skills to non-English users. Program: Electronic Library and Information Systems, 41 (2), 170–179. [8] Chen, L. S., & Chen, S. L. (2007). Collaborative design and manufacture on intelligent system. Journal of the Chinese Society of Mechanical Engineers, 28 (2), 233–242. [9] Duan, W., Gu, B., & Whinston, A. B. (2009). Informational cascades and software adoption on the internet: an empirical investigation. MIS Quarterly, 33 (1), 23–48. [10] Abbasi, A., & Chen, H. (2008). CyberGate: a design framework and system for text analysis of computer mediated communication. MIS Quarterly, 32 (4), 811–837. [11] Feller, J., Finnegan, P., Fitzgerald, B., & Hayes, J. (2008). From peer production to productization: a study of socially enabled business exchanges in open source service networks. Information Systems Research, 19 (4), 475–493. [12] Agarwal, R., Gupta, A. K., & Kraut, R. (2008). Editorial overview—the interplay between digital and social networks. Information Systems Research, 19 (3), 243–252.

10

L.-S. Chen / Library Collections, Acquisitions, & Technical Services 34 (2010) 1–10

[13] Bapna, R., Goes, P. B., & Gupta, A. (2009). Auctioning vertically integrated online services: computational approaches for real-time allocation. Journal of Management Information Systems, 25 (3), 65–98. [14] Hussain, Z. I., & Cornelius, N. (2009). The use of domination and legitimation in information systems implementation. Information Systems Journal, 19 (2), 197–224. [15] Larsen, T. J., Niederman, F., Limayem, M., & Chan, J. (2009). The role of modelling in achieving information systems success: UML to the rescue. Information Systems Journal, 19 (1), 83–117. [16] Gilstrap, D. L. (2009). A complex systems framework for research on leadership and organizational dynamics in academic libraries. Portal: Libraries and the Academy, 9 (1), 57–77. [17] Saunders, L. (2009). The future of information literacy in academic libraries: a Delphi Study. Portal: Libraries and the Academy, 9 (1), 99–114. [18] Saunders, E. S. (2008). Meeting academic needs for information: a customer service approach. Portal: Libraries and the Academy, 8 (4), 357–371. [19] Dana, H. B., & Sandra, H. H. (2009). The information and communication technology competencies of students enrolled in school library media certification programs. Library & Information Science Research, 31 (1), 3–11. [20] Aharony, N. (2009). Web 2.0 use by librarians. Library & Information Science Research, 31 (1), 29–37. [21] Chung, H. K. (2009). An analysis model of creating a core journal collection for academic libraries. Library Collections, Acquisitions, and Technical Services, 33 (1), 17–24. [22] Pomerantz, S., & White, A. (2009). Re-modeling ILS acquisitions data to financially transition from print to digital formats. Library Collections, Acquisitions, and Technical Services, 33 (1), 42–49. [23] Chen, L. S. (2008). Design and implementation of intelligent library system. Library Collections, Acquisitions, and Technical Services, 33 (3-4), 127–141. [24] Duinkerken, W., Smith, J., Harrell, J., Reynolds, L. J., Tucker, S., & Carrigan, E. (2008). Creating a flexible fund structure to meet the needs and goals of the library and its users. Library Collections, Acquisitions, and Technical Services, 33 (3-4), 142–149. [25] Adkins, D., Esser, L., Velasquez, D., & Hill, H. L. (2008). Romance novels in American public libraries: a study of collection development practices. Library Collections, Acquisitions, and Technical Services, 33 (2), 59–67. [26] Termens, M. (2008). Looking below the surface: the use of electronic journals by the members of a library consortium. Library Collections, Acquisitions, and Technical Services, 33 (2), 76–85. [27] Quinn, B. (2008). The psychology of group decision making in collection development. Library Collections, Acquisitions, and Technical Services, 33 (1), 10–18. [28] Martinovic, G., & Cukic, B. (2008). Multicomputer system for optical character recognition in Web library creation. Library Collections, Acquisitions, and Technical Services, 33 (1), 19–30. [29] Martens, D., De Backer, M., Haesen, R., Vanthienen, J., Snoeck, M., & Baesens, B. (2007). Classification with ant colony optimization. IEEE Transactions on Evolutionary Computation, 11 (5), 651–665. [30] Birattari, M., Pellegrini, P., & Dorigo, M. (2007). On the invariance of ant colony optimization. IEEE Transactions on Evolutionary Computation, 11 (6), 732–742. [31] Wang, T. I., Wang, K. T., & Huang, Y. M. (2008). Using a style-based ant colony system for adaptive learning. Expert Systems with Applications, 34 (4), 2449–2464. [32] Nolle, L. (2008). On a novel ACO-estimator and its application to the target motion analysis problem. Knowledge-Based Systems, 21 (3), 225–231. [33] Marinakis, Y., & Dounias, G. (2008). Nature inspired intelligence in medicine: ant colony optimization for pap-smear diagnosis. International Journal on Artificial Intelligence Tools, 17 (2), 279–301. [34] Tewolde, G.. S., & Sheng, W. (2008). Robot path integration in manufacturing processes: genetic algorithm versus ant colony optimization. IEEE Transactions on Systems, 38 (2), 278–287.