Effectiveness of Learning Transportation Network Growth through Simulation
Wenling Chen and David Levinson
Department of Civil Engineering
University of Minnesota, Twin Cities
500 Pillsbury Dr. SE
Minneapolis, MN 55455
Draft July 29, 2004
Abstract
Computer simulation plays an increasingly important role in engineering education as a tool for enhancing classroom learning. This research investigates the efficacy of using simulation in teaching the topic of transportation network growth through an experiment conducted at the Civil Engineering Department of the University of Minnesota. In the experiment, a network growth simulator program (SONG 1.0) was incorporated into a senior/graduate class in transportation system analysis. Results of the experiment show that the use of SONG 1.0 effectively enhanced students learning in terms of helping students develop in-depth understanding about the development process of network patterns, and helped them develop some aspects of judgment, problem-solving, and decision-making skills.
Key words:
Simulation, Engineering Education, and Transportation Network Growth
Introduction
Conventional transportation engineering and planning education addresses the topic of travel demand modeling and network growth dynamics through lectures on general theories, pencil and paper problems, or class projects on related topics. Although this traditional approach imparts knowledge, there remains room to explore alternative teaching strategies to improve teaching outcomes by accommodating different learning styles, promoting active learning, and providing opportunities for students to practice judgment and problem-solving skills.
Simulation complements traditional engineering education methods. Simulations are engaging and allow learners to internalize knowledge by applying new skills in a risk-free environment, which can dramatically increase motivation and retention rates and provide a high return on learning efforts (Billhardt, 2004). Despite its potential, simulations are used infrequently in transportation engineering and planning education. One major barrier that prevents wide adoption of simulation lies in the uncertainty over how to develop, use, and incorporate simulations successfully into existing education environment (Billhardt, 2004).
To bridge the gap, a transportation network growth simulator is developed and incorporated into a senior/graduate level transportation system analysis course as an assignment. Based on the assignment, the authors designed an experiment which enables an efficacy test on the network growth simulator as used in this class. This paper documents the process of the experiment and reports the findings from the evaluation.
Simulation in Education
Advanced education and teaching are increasingly based on technological innovations in the area of multimedia and computer-based instruction (Alavi et al., 1997). One of these innovations is the application of simulation. By definition, simulation is a dynamic representation of some part of the real world by building a computer model and moving it through time (Drew 1968). Simulation allows learners to engage actively by running experiments, testing different strategies, and building a better understanding of the aspects of the real world which the simulator depicts (Pursula, 1999). In simulation, learners' individual choices lead them down different paths toward different outcomes. Essentially, simulation lets students learn directly from the outcomes of their own actions (Senge, 1990; Billhardt, 2004; Aldrich, 2003). In particular, the value of simulation in transportation education can be summarized as the following:
Simulation provides learners with experiences: The importance of experience in human learning has long been emphasized. Phenomenological studies of human learning indicate that people pass through several levels in the learning of skills, ranging from the technical to the intellectual. High-level performance within a given area requires expertise based on experience, intuition, and judgment (Dreyfus and Dreyfus, 1986; Flyvbjerg, 2001). Conventional approaches to transportation education emphasize rationality and are dominated by analytical training, which tend to deemphasize sensitivity to experience, context, and intuition (Flyvbjerg, 2001). One reason for less emphasis on experiential learning lies in the fact that real world experience in transportation is difficult to apply to classroom learning, because effects of transportation policies take decades to materialize; additionally, the risks and costs of experimenting transportation policies and concepts in the real world are prohibitively high. Simulations compress time and space. Through simulation, experiential learning can be facilitated and encouraged.
Simulation provides opportunities for learning through doing : many people learn best through taking actions, or learning by doing (Lowman, 1984; Mckeachie, 1986; Senge, 1990; Dreyfus 1986; and Lerman, 2002). Rationales for learning by doing are rooted from the constructivist learning theories of Jean Piaget (1955), according to whom, knowledge is constructed, discovered, transformed, and extended by learners; the role of faculty is to create condition to facilitate knowledge construction by students (Johnson, et, al., 1998; Lyons, 2001). Simulation creates an environment to engage students in experiments and knowledge construction (Resnick, 1997).
Simulation provides interactive learning environment: many students also learn from experience, but this learning only occurs if the consequences of actions and decisions are experienced in a rapid and unambiguous manner (Senge, 1990; Billhardt, 2004). Providing quick feedback in an interactive manner is one of the advantages of simulators compared to other tools of experiential learning such as case studies, which, while allowing students to experience decision-making, are less effective in providing feedback. In simulation, feedback can be given right after an action is taken, in which way learners tend to be more open to internalizing knowledge (Billhardt, 2004).
Simulation diversifies teaching strategies: Diversifying teaching methods helps learning because it is a way to accommodate students' different learning styles. Research shows that no single learning style leads to better learning, however, benefits of certain teaching strategies can only be caught by students with certain learning styles; teaching while meeting different learning styles and orientations enhance teaching effectiveness (Cross, 1976; Matthews, 1991; Davis, 1993; Kolb, 1984; Perry, 1970).
Simulation helps students move toward higher levels of intellectual development: Human learning develops with cognitive development ranging from feeling, watching, and thinking to doing (Kolb, 1984). In terms of classroom activities, simulation has been identified as most suitable for students to develop and practice the highest stage of intellectual growth (Svnicki and Dixon, 1987; Claxton and Murrell, 1987; Erickson and Strommer, 1991; Fuhrmann and Grasha, 1983).
Simulation engages motivation to learn: Effective learning in the classroom depends on the teacher' s ability to motivate students and maintain their interests to participate in the course in the first place (Ericksen, 1978, p.3). General strategies of motivating students include actively involving students to learn through doing, and vary teaching methods to reawaken students' involvement in courses (Forsyth and McMillan, 1991), all of which can be achieved through the use of simulation.
Simulator of Network Growth (SONG 1.0)
The Simulator of Network Growth (SONG 1.0), which can be accessed at http://www.ce.umn.edu/~levinson/Song/Dynamics.html, supports the learning of the transportation network development process. The growth or decline of transportation networks is normally treated as the result of top-down decision making in long-range planning efforts of metropolitan planning organizations (MPOs). However, changes to transportation networks are essentially the result of numerous small decisions by property owners, firms, developers, towns, cities, counties, state department of transportation districts, MPOs, and states in response to market conditions and policy initiatives (Yerra and Levinson, 2004). This kind of system behavior demonstrates the characteristics of decentralized systems, where organized patterns and structures can emerge not because of centralized control, but because of the interactions among decentralized system components. In SONG 1.0, transportation networks are treated as decentralized systems that demonstrate the property of self-organization. The simulator models behaviors of individual system components (network links) and small decisions, and then demonstrates the patterns resulting from interactions among the component models.
As illustrated in the modeling process of SONG 1.0 (Figure 1), SONG 1.0 treats each network link as an autonomous agent. The program takes exogenous inputs such as the base network and land use distribution, and translates them into traffic flows and speeds on network links through travel demand model. Those traffic flows and speeds determine the revenue and costs of maintaining and improving the link, and inform the network investment model. When each link has exhausted its resources, the time period is incremented, population grows, land uses are updated, the travel demand is recomputed on the new network, and the process repeats. At the end of the process, data is exported to a visualization tool, which will allow the growth to be seen in a movie-like fashion (Yerra and Levinson, 2004).
In the interface of SONG 1.0, as shown in Figure 2, users can adjust parameters to change travelers' value of time, their willingness to travel, toll, how revenue and cost change in response to changes in road speed, flow and distance traveled and how investments are determined based on link performance. By adjusting these parameters, users can test the effects of these factors on the resulting network forms, which are visualized in terms of speeds or volumes on network links represented by different colors and thickness of the links.
An example, shown in Figure 3, illustrates how SONG 1.0 works. Figure 3 demonstrates different network patterns evolved from different elasticity of link maintenance costs to speed change. The initial network is shown on Figure 3a, where there are no speed differences across links. Figure 3b displays the resulting network speed pattern with cost elasticity adjusted such that a one percent increase in speed will lead to less than one percent increase in road maintenance costs, indicating an economy of scale in upgrading road speed. Figure 3c shows the resulting network speed pattern with a diseconomy of scale in upgrading road speeds holding all other factors constant.
Users can draw two implications from the simulation: first, hierarchical patterns emerge out of a uniformly laid out network with fewer higher-speed links clustered around the center and a larger number of lower-speed links distributed adjacent to the network borders; and second, an economy of scale (Figure 3b) leads to higher level of investment to increase road speed, and diseconomy of scale to speed upgrade creates less incentives for speed upgrade, resulting in generally lower and more uniform speeds across the network (Figure 3c).
The same experiments can be done in a randomized manner as demonstrated in Figure 3d-f, where similar network patterns occur except that randomized speed distribution leads to more stochastic network patterns instead of the symmetric patterns shown in Figure 3a-c.
Applied in educational settings, SONG 1.0 is expected to stimulate students to think and gain new understandings about how transportation networks grow or decline, whether network patterns (e.g., hierarchy) are planned or emergent, and how policy alternatives affect the location of network expansions and contractions. Compared with other simulators or the software packages commonly used in transportation education, SONG 1.0 has the following features:
Soft Simulation: SONG 1.0 is a soft simulation , which provides a qualitative understanding of a complex system by constructing a simple one that shares the same principle (Papert, 1992). In many cases, simulations are designed to imitate and make predictions about real-world systems as accurately as possible. However, in SONG 1.0, more interest is placed on stimulation than in simulation . In developing the simulator for classroom use, the focus is not on a perfect reproduction of the real world, but rather to help students explore the microworld of transportation network systems and to stimulate new ways of thinking about the network growth and planning process.
Simpler, Easier, and Cheaper: Conventional planning software packages, such as EMME/2, TransCAD, and TranPlan, are often cumbersome, difficult to learn, and expensive. SONG 1.0 is simpler and easier to learn, and is free for students to use. Hence, it costs both students and instructors less to incorporate SONG 1.0 into the curriculum.
Network Growth Model: SONG 1.0 is also distinguished from other transportation simulation programs in that it is a network growth model. So far, the authors have discovered no literature on educational application of network growth models.
Given its features, SONG 1.0 is expected to have a value in the teaching of transportation network evolution. This study investigates the usefulness and efficacy of SONG 1.0 as an educational tool by adopting SONG 1.0 into a transportation planning/ engineering course.
Experiment
To investigate whether or not the use of SONG 1.0 would enhance the learning of transportation network growth, an experiment was conducted on a senior/graduate level course on transportation systems analysis in spring semester 2004 at the Civil Engineering Department, University of Minnesota. The objectives for adopting SONG 1.0 to classroom education are threefold:
Improve Learning Outcomes: Besides learning the subject of transportation network growth, SONG 1.0 is also expected to help students develop soft skills in judgment and problem-solving through the experiential learning obtained in the simulated environment. The particular learning outcomes expected through using SONG 1.0 include:
- Stimulate new ways of thinking about the dynamics of network development
- Enhance the ability to draw implications of alternative policies on transportation network form
- Develop understanding of the transportation network development process, the influencing factors and players
- Develop an understanding of travel demand modeling process
- Develop problem-solving skills and judgment skills in infrastructure investment decision making
Test Hypothesis: The research objective of this experiment is to investigate if the use of simulation can improve the learning outcomes, and test the hypothesis that SONG 1.0 can be an effective tool for enhancing students' learning on the subject of transportation network growth.
Generate Guidelines for Applying Simulation in Transportation Education: Experience, findings and lessons learned from this study will be summarized to provide implementation guidelines for attempts to innovate in teaching through the use of simulation.
Transportation Systems Analysis (CE 5214) is a 3-credit senior/graduate course. The course objectives are to have students acquire knowledge of travel behavior, travel demand forecasting, and network growth, and to develop context sensitive use of problem-solving and judgment skills necessary for success in the transportation profession as civil engineers and planners. Previously, the teaching is based on traditional approaches of lecturing, problems, and examinations. While these approaches led to learning by students, they may not fully foster learning and application of knowledge. Hence, SONG 1.0 is incorporated into this course as an innovation for improving teaching effectiveness. Table 3 summarizes the students' backgrounds.
The experiment incorporates SONG 1.0 into a homework assignment of CE 5214. The experiment process, shown in Figure 4, contains three parts: 1) Comparative Study on Two Groups: the experiment randomly divided the class into two groups with the control group receiving the traditional case study-based assignment (see Appendix B), and the treatment group taking SONG 1.0-based assignment (see Appendix A). A comparative study on the two groups is to determine whether students learn better with SONG 1.0 than without it; 2) Assignment Design: As shown in Table 1, the control assignment and treatment assignment are designed such that the objectives, substances, and work loads are of no significant difference. They differ, however, in that the treatment assignment is based on simulation platform, it allows students to make changes and see consequences of their actions, it allows students to see the visualized outcomes, it is interactive, and allows students to learn through doing ; and 3) Evaluation: Two surveys and one exam are conducted to assess students' performances and investigate SONG 1.0' s efficacy as a learning tool.
Evaluation
As shown in Table 2, the evaluation involves two steps: control students' background differences and other confounding factors, and compare learning outcomes between the two groups. Data for the evaluations are collected from the pre-assignment survey (see Appendix C), the post-treatment survey (see Appendix D), and final exam (see Appendix E).
Beyond the simulator, many other factors also affect students' learning. An analysis of these factors provides critical information for determining whether the differences (or indifferences) in learning outcomes can be attributed to the effects of the simulator. In particular, students' academic background, relevant prior experiences and knowledge, computer proficiency, and learning styles are expected to affect their performance in the assignment.
In this study, self-reported learning styles are assessed with Kolb (1984)' s Learning Style Inventory (LSI), and Felder-Silverman (1988)' s Index of Learning Styles (ILS). As shown in Figure 5, LSI is an established tool for learning style assessment; while ILS is developed mainly to assess learning styles of engineering students (Evans, et al., 2000). It is expected the educational benefits of SONG 1.0 are most likely to be captured by students with preferences to learn through watching, and doing, and students who prefer visual and active styles of learning.
With students' background differences and other confounding factors being controlled, students' learning outcomes are compared to determine whether the use of SONG 1.0 leads to different learning by the two groups. Learning outcomes are measured with three criteria: (1) time taken to complete the assignment, (2) achievement of learning objectives, including subject understanding, and skills improvement, and (3) students' reflections on the learning experiences with the assignments.
Of 31 students, 27 students finished the homework on time, 26 students responded to the pre-treatment survey, and 25 students responded to the post-treatment survey. Results of the assessment on students' background, learning styles, prior knowledge, and prior skills are shown in Table 3.
Demographic, Academic, Professional Background and Technical Capacity Assessment indicates that none of the factors examined are significantly different between the treatment group and the control group; learning style assessment shows no significant differences between the two groups either. It is also revealed that 15 out of 28 students prefer learning through watching and 20 out of 28 students prefer learning through doing , implying that the use of simulator matches the learning preferences of the majority of the class; Prior knowledge assessment indicates that the treatment group is significantly less familiar with travel demand modeling process, and transportation simulation; and Prior Skills Assessment decomposes and evaluates students' judgment skills and problem-solving skills at a factor level. It is implied that the control group perceives themselves significantly stronger in terms of forming opinions (judgment skill) and developing methods to solve problems (problem-solving skill). The same differences persisted through post-treatment survey.
Learning outcomes are measured in terms of students' performance, time spent on the assignment, and students' reflections on their assignment learning experience. Student performance is assessed both through surveys in terms of their perceived improvements on skills and subject understanding, and through their performance on the final exam.
Three exam questions test the subjects of travel demand modeling process, network development process and students' problem-solving skills in infrastructure investment decision-making. To assess students' ability to apply the concept learned, students were asked to use examples to illustrate their answers.
The exam grading criteria include: relevance of the answers how closely and clearly the questions were addressed, application of the concept how well examples are interpreted, and depth of understanding on the subject examined. Depth of learning is assessed in terms of understanding, understanding the subject in a different ways, and incorporating learners' own position and perspectives (Romme, 2002)
The reason for evaluating students' depth of understanding is that the treatment group is expected to lean toward deeper learning than the control group. Different from surface learning, which is tied to a specific learning situation given, such as a text, problem or assignment (Martin, 1999; Romme, 2002), a deep learning goes beyond the given situation or problem, and explores the larger issues represented by a particular problem (Martin, 1999). SONG 1.0 is expected to be more productive and valuable in facilitating deep learning because of the interactive situation and complex interplay of variables provided through the simulation. Results of learning outcome assessment are summarized in Table 4.
Performance assessed through surveys: In terms of students' perceived improvements on subject understanding through the assignment, the treatment group enhanced their understanding significantly better than the control group about development process of network pattern. In terms of skill improvements assessed by comparing perceived skill changes through the assignment between the two groups, the treatment group indicated significantly more improvements than the control group in terms of their ability to identify the relationship of components in transportation systems and the ability to use established criteria to evaluate and prioritize solutions.
Performance Assessed through Final Exam Questions: In terms of subject understanding, which were assessed through questions on the four-step travel demand modeling, and on development process of network pattern, the treatment and control groups were found to perform equally well. In terms of students' decision-making and problem-solving skills, the overall performance of the treatment group is found to be significantly better than the control group.
Time spent on completing the assignment and students' reflections on the experiences of learning through the experimental homework is another aspect of performance being examined. After the experiment, students were surveyed about the time they spent on completing the assignment, their satisfaction with the amount of time they have spent, as well as their effectiveness in completing the assignment. No significant differences between the treatment and control groups were found in these regards. Students' reflections on the learning experiences, which are informative as for how well they learned and how helpful was the teaching strategy they experienced, are of no significant differences between the two groups either.
Learning Outcomes vs. Students' Characteristics: To explore what kind of students gain most from the simulation-based assignment, regressions were run between students' performances and several explanatory variables on students' characteristics. As shown in Table 5, taking simulation-based assignment does positively and significantly associate with students self-reported improvements in their understanding of the development of network patterns. Contrary to expectation, students who prefer learning through thinking (reflective learning) instead of learning through doing (active learning) are positively associated with understanding improvements on this subject with statistical significance. At the 99% confidence level, simulation treatment is positively associated with students' perceived improvements on the ability to identify relationships of components in transportation systems. Students' relevant working experience is negatively correlated with improvement on this particular skill.
Students' performance on the exam is also associated with some of their characteristics as shown in Table 5. In terms of the question on network development pattern, students' ability to incorporate their own perspective into the answers is positively related to their age and relevant working experiences. Students who are more oriented to global and holistic thinking and those who prefer constructing their own knowledge are more likely to perform well on this regard. In terms of students' decision-making skills, it is found that, students with more relevant courses taken before are more likely to perform better on this question; and students with preference to constructing their own knowledge and who have taken more relevant courses showed stronger ability to understand the subject in different ways.
Conclusions and Lessons Learned
Findings from this research can be summarized as the follow: first, the use of SONG 1.0 is effective in improving students' performance in some areas of learning. With SONG 1.0, students performed significantly better in learning network development patterns and in developing their ability to identify a relationship of components in transportation systems, the ability to establish criteria to evaluate and prioritize solutions, in developing decision-making skills and in-depth understanding of the investment decision making process.
Second, as summarized in Table 6, those who performed better in certain learning areas possess certain characteristics in terms of their age, education level, computer proficiency, prior experience as well as learning styles. Hence, for different learning outcomes pursued, it can be effective to apply simulation to learners of appropriate age, educational level, learning styles and prior knowledge;
Third, for most of the learning outcomes assessed, the treatment group performed equally well as the control group. As revealed from the surveys, the lower-than-expected learning outcomes achieved by the treatment group can be explained by three factors: (1) in terms of prior knowledge and skills, the control group had significant advantages over the treatment group as indicated in the background assessment; (2) design of SONG 1.0: a good educational simulator depends on its complexity and feedback (Billhardt, 2004), while this study indicated that the messages SONG 1.0 sent were not clear or self-explanatory to the treatment group. It is also indicated that SONG 1.0 was not complex enough to incorporate some of the real world situations students were interested to test; and (3) in terms of course design, insufficiency of instruction and supporting information, as well as lack of clarity in instruction and supporting information were indicated by the treatment group as problems for better learning; additionally, timing and workload were indicated as problems since the assignment was near the semester end, when students are likely to be overloaded, introducing pressure as another confounding factor to this study.
Issues with course design and SONG 1.0' s usability created a barrier to learning, and prevented students from capturing the full educational benefits of SONG 1.0. This provides valuable lessons for guiding future practice in adopting simulation into educational setting:
- Provide reasonable complexity of simulator
- Feedback from simulators needs to be unambiguous and self-explanatory
- Interactive instruction is desirable: To help students learn simulators, interactive lab instruction is more effective in removing technical barriers than one-way lecturing.
- Proper work load and timing: Work load and timing of the assignment needs to be carefully designed so that students can be given more time to play , and the fun of simulation is more likely to materialize.
- Clear assignment instruction: Be specific about the tasks students need to fulfill the assignments. This creates effective orientation of the substance students are expected to learn from the simulator.
- Maintain sufficiency of instructions and supporting information
References
1. Alavi, M., Y. Yoo, et al. Using information technology to add value to management education. Academy of Management Journal 40(6) 1997, pp. 1310-1333.
2. Aldrich, Clark. Simulations and the Future of Learning: An Innovative (and Perhaps Revolutionary) Approach to e-Learning. Pfeiffer, 2003.
3. Billhardt, Bjorn (2004). The Promise of On-line Simulations. MediaTec Publishing Inc. Assessed February 2004 <http://www.clomedia.com/content/templates/clo_feature.asp?articleid=382&zoneid=29>.
4. Claxton, Charles S. and Patricia H. Murrell (editors). Learning styles: Implications for improving education practices. ASHE-ERIC Higher Education Report, 4. Washington: Association for the study of Higher Education. 1987, pp. 79.
5. Cross, K.P. Accent on learning: Improving instruction and reshaping curriculum. San Francisco, CA: Jossey Bass, 1976.
6. Davis, Barbara Gross. Tools for Teaching. Jossey-Bass, 1993.
7. Dreyfus L, Dreyfus, S E. Mind over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. Oxford: Basil Blackwell, 1986.
8. Drew, D.R. Traffic Flow Theory and Control. McGraw-Hill, New York. Ericksen, 1968.
9. Erickson, B. L., and Strommer, D. W. Teaching College Freshmen. San Francisco: Jossey-Bass, 1991.
10. Evans, R.M., Murray, S.L., Daily, M., & Hall, R.H. Effectiveness of an Internet-Based Graduate Engineering Management Course. Journal of Engineering Education, 89(1), 2000, pp. 63 71.
11. Felder, R.M. and Silverman, L.K. Learning and Teaching Styles in Engineering
Education. Engineering Education. 78(7), 1988, pp. 674-681.
12. Flyvbjerg, Bent. Making social science matter: why social inquiry fails and how it can succeed again. New York: Cambridge University Press, 2001.
13. Forsyth, D. R. and McMillan, J. H. Why do learners learn? Answers offered by current theories of motivation. In R. J. Menges & M. Svinicki (Eds.), Approaching Instructional Problems through Theoretical Perspectives: New Directions for teaching and Learning. San Francisco: Jossey-Bass, 1991, 44, pp. 39-52.
14. Fuhrmann, B. S., and Grasha, A. F. A Practical Handbook for College Teachers. Boston: Little, Brown, 1983.
15. Johnson, David W., Johnson, Roger T., and Smith, Karl A. Active Learning: Cooperation in the College Classroom, 1991, 1998.
16. Kolb, D.A. Experiential Learning: Experience as the Source of Learning and Development. New Jersey: Prentice-Hall Inc., 1984.
17. Lerman, Steven R. (2002) Some Criteria for the Evaluation of Multimedia Computer Applications. Assessed April 2004 <http://design.chosun.ac.kr/ddl/hdb/data/mul/lerman2.htm>.
18. Lowman, J. Mastering the techniques of teaching. San Francisco: Jossey-Bass. Lyons, 2001.
19. Martin, E. Changing Academic Work. Buckingham: Open University Press, 1999.
20. Matthews, D. B. The effects of learning styles on grades of first-year college students. Research in Higher Education, 32(3), 1991, pp. 253-268.
21. McKeachie, W. Meeting a Class for the First Time. In Teaching Tips: A Guidebook for the Beginning College Teacher, 8th ed. Lexington, MA: D.C. Heath and Company, 1986, pp. 18-24.
22. Papert, Seymour, The Children' s Machine: Rethinking School in the Age of the Computer, 1992.
23. Perry, W. Forms of intellectual and ethical development in the college years: A scheme. New York: Holt, Reinhart & Winston, 1970.
24. Piaget, Jean. The Construction of Reality in the Child, Routledge and Kegan Paul, 1955.
25. Pursula, Matti. Simulation of traffic Systems-An Overview. Journal of Geographic Information and Decision Analysis vol.3, no.1, 1999, pp.1-8.
26. Resnick, M. Turtles, Termites, and Traffic Jams: Explorations in Massively Parallel Microworlds. Cambridge, MA: The MIT Press, 1997.
27.
Romme, Georges L. Microworlds for Management Education and Learning. Tilburg,
28. Senge, Peter M. The Fifth Discipline: The Art and Practice of the Learning Organization. New York: Doubleday Currency, 1990.
29. Svnicki, Marilla D. and Dixon, Nancy M. The Kolb Model Modified for Classroom Activities. College Teaching, Vol. 35, No. 4 Fall, 1987, pp. 141-146.
30. Yerra, B. and D. Levinson 2004, The Emergence of Hierarchy in Transportation Networks (How Networks Grow). 2004.
List of Tables and Figures
Table 1: Comparison of Simulation-Based Assignment vs. Case Study-Based Assignment
Table 2: Evaluation Design
Table 3: Assessment on Students' Background, Learning Styles, Prior Subject
Understanding, and Prior Skill
Table 4: Learning Outcome Assessment
Table 5: Perceived Improvements of Understanding and Skills vs. Students'
Characteristics & Exam Performance vs. Students' Characteristics
Table 6: Summary of Characteristics of Students and Learning Outcomes Achieved with
Treatment Group Performing Significantly Better
Figure 1: Modeling Process Flowchart of SONG 1.0
Figure 2: Interface of SONG 1.0
Figure 3: Effects of cost elasticity to speed changes (uniform vs. randomized patterns)
Figure 4: Experiment Design
Figure 5: Kolb Learning Style Inventory (LSI) and Felder-Silverman Index of
Learning Style (ILS)
Table 1: Comparison of Simulation-Based Assignment vs. Case Study-Based Assignment
Table 2: Evaluation Design
Table 3: Assessment on Students' Background, Learning Styles, Prior Subject Understanding, and Prior Skill
* Significant at 90% confidence level
** Significant at 95% confidence level
*** Significant at 99% confidence level
Table 4: Learning Outcome Assessment
Table 5: Perceived Improvements of Understanding and Skills vs. Students'
Characteristics & Exam Performance vs. Students' Characteristics
Table 6: Summary of Characteristics of Students and Learning Outcomes Achieved
with Treatment Group Performing Significantly Better
Figure 1: Modeling Process Flowchart of SONG 1.0
Figure 2: Interface of SONG 1.0
Source: SONG 1.0 Help file, 2004
Figure 3: Effects of cost elasticity to speed changes (uniform vs. randomized pattern)
Figure 3a: Initial speed distribution (uniform) |
Figure 3d: Initial speed distribution (random) |
|
Figure 3b: Speed reach equilibrium with economy of scale in speed upgrade (uniform) |
Figure 3e: Speed reach equilibrium with economy of scale in speed upgrade (random) |
|
Figure 3c: Speed reach equilibrium with diseconomy of scale in speed upgrade (uniform) |
Figure 3f: Speed reach equilibrium with diseconomy of scale in speed upgrade (random) |
|
Figure 4: Experiment Design
Figure 5: Kolb Learning Style Inventory (LSI) and Felder-Silverman Index of
Learning Style (ILS)
Source:
Kolb, 1976; Felder and Silverman, 1988; & Evans et al., 2000
Appendices:
Appendix A: Assignment for treatment group
CE 5214 Transportation System Analysis (Spring 2004)
Homework 4A
Transportation Network Development System Analysis
Due on: April 27, 2004
The main objectives of this assignment are to help student
- learn to draw implications of alternative policies on transportation network
- understand transportation network development process, the influencing factors and players
- learn transportation planning model (modeling process)
- understand current transportation infrastructure investment decision making process
- learn and practice the method of system analysis
Statewide transportation plan is one of the major products generated out of the federally-mandated transportation planning process. The transportation plan is a long range (at least 20 years), multi-model future vision for the mobility of goods and people. The plan considers factors that may affect or be affected by local and regional transportation investments. Minnesota Department of Transportation (MnDOT) is the agency holding responsibility for administrating the state transportation budget. Currently, MnDOT is drafting the 2004 to 2024 Minnesota Statewide Transportation Plan. In preparation for the statewide transportation plan, MnDOT is developing a 20-year transportation investment and financing strategy for the development of transportation system.
You are a transportation planner/engineer working for a local transportation consulting firm, which newly won a contract with MnDOT to perform a pilot study to test in concept the effects of possible policy initiatives or alternative decision-making assumptions in transportation investment and financing. In particular, your firm is contracted to explore the implications of the following changes in investment policy and decision-making assumptions:
(1) Changes of assumptions in travel behaviors, such as:
· Value of time
· Willingness to travel
(2) Changes in toll charged for using roads
(3) Changes in revenue elasticity in response to distance traveled, and road standard, e.g., link speed
(4) Changes in elasticity of road maintenance costs in response to road length, flow, and speed
(5) Changes in level of investments based on link performance
Additionally, MnDOT requests that these changes be examined under different network and land use contexts such as:
(1) Different land use density
(2) Different network speed levels
For this project, your firm purchased a network growth simulation software lately developed, called Simulator of Network Growth (SONG 1.0), which has been made available for this course under the following link: http://www.ce.umn.edu/~levinson/Song/Dynamics.html. Your supervisor, Ms. Nelson, assigned you to work on this project, and in two week, you have to get yourself familiar with the software, perform system analysis on the network alternatives under different policy scenarios and decision-making assumptions that MnDOT is interested in, and you must submit a memo to report your findings.
The simulation tool: SONG 1.0
A 30 minutes' instruction will be given to help you learn how to use and understand the underlying model of SONG 1.0. For your brief information, SONG 1.0 is a network growth simulator, developed based on the traditional four-step transportation planning model. It is designed to visualize and enhance better understanding of how transportation networks grow. The policy scenarios MnDOT is interested in have been customized into the software. Additionally, SONG 1.0 contains other policy variables that allow you to adjust to reflect different land use and network situations. Outputs of the simulation include graphical representations of the network volumes and speeds under equilibrium, and a set of system measurements of effectiveness (MOEs) for further system analysis. In particular, the following MOEs can be generated from the simulation:
- Average speed
- Average volume
- Vehicle Kilometer Traveled (vkt)
- Vehicle Hours Traveled (vht)
- Total cost
- Total revenue
- Cumulative costs
- Cumulative revenue
- Improvement term
Your Tasks
In completing this project, you must fulfill the following tasks:
Task 1: Understand the simulator
Run simulations under default values as well as two of the remaining sets of land use-speed values and interpret the results:
· Base case (default values: uniform land use, uniform speed)
· Uniform land use, random speed
· Random land use, uniform speed
· Random land use, random speed
· Downtown land use, uniform speed
· Downtown land use, random speed
Task 2: Run the simulation under different policy scenarios of interest. (You can adjust values of parameters to reflect different policies or assumptions). Copy the graphic output for your report. (You can use copy screen function of you computer, and then paste the screen with graphic output in a world file).
Task 3: Perform system analysis based on the MOEs output from the simulation. As an example, a system analysis can contain the following steps:
1) Define the system (e.g., network types, speed, and land use distribution)
2) Generate and assess alternatives available (policy scenarios of interest)
3) Choose alternatives and implement (run simulations)
4) Get feedback (compute and summarize the MOEs)
5) Evaluate and select preferred policy scenarios.
Evaluation can be made based on comparison of the MOEs on certain criteria, for example (you' re NOT required to use all of the following criteria):
a. Effectiveness and Efficiency
· Mobility: Travel time (vht)
· Accessibility: Delay, access to desired locations, access to system
· Reliability: variability of travel time
· Cost-effectiveness: benefit/ cost ratio, outcome benefit per costs
· Consumer surplus
b. Responsibility
· Sustainability: transportation costs
· environmental quality: national/state standard
· Safety and security: accident and crime rate
· Equity: benefit per income group
· customer satisfaction
· economic well being
Also, the following issues should be considered when making the comparisons: scale vs. detail, time frame, boundary effects, short term vs. long term, centralized vs. decentralized, and etc.
Besides method of systems analysis, other analysis methods are also available. You can select analysis methods on your own based on your understanding of how the evaluation should be performed and which methods are applicable to the issues of interest. Examples of other analysis methods include:
· Network analysis model
· Cost/benefit analysis model
· system dynamics model
· probabilistic risk assessment
· Statistical decision theory
· etc.
Task 4: Submit a memo to your supervisor and report your findings.
Recommended outline for your report is as the follow:
1) Problem statement
2) Methodology
- Simulation (briefly describe SONG 1.0 and report your results from Task 1)
- Analysis methodology (stating what and why you choose a particular method)
3) Evaluation and Analysis
4) Results and Findings
5) Discussion of limitations
6) Conclusion
The report must be no more than 2500 words, 1.5 or double spaced, using 12-point Times New Roman font.
Appendix B: Assignment for the Control Group
CE 5214 Transportation System Analysis (Spring 2004)
Homework 4B
Transportation Network Development System Analysis
Due on: April 27, 2004
The main objectives of this assignment are to help student
- learn to draw implications of alternative policies on transportation network
- understand transportation network development process, the influencing factors and players
- learn transportation planning model (modeling process)
- understand current transportation infrastructure investment decision making process
- learn and practice the method of system analysis
Statewide transportation plan is one of the major products generated out of the federally-mandated transportation planning process. The transportation plan is a long range (at least 20 years), multi-model future vision for the mobility of goods and people. The plan considers factors that may affect or be affected by local and regional transportation investments. Minnesota Department of Transportation (MnDOT) is the agency holding responsibility for administrating state transportation budget. Currently, MnDOT is drafting the 2004 to 2024 Minnesota Statewide Transportation Plan. In preparation for the statewide transportation plan, MnDOT is developing a 20-year transportation management and financing strategy for the development of transportation system.
You are a transportation planner/engineer working for a local transportation consulting firm, which newly won a contract with MnDOT to perform a pilot study to examine potential system impacts of statewide application of a transportation management policy: High Occupancy Vehicle (HOV) lanes as opposed to the general purpose lanes. The project team of your firm decides to phase this study into two steps: (1) conduct a corridor level impact analysis on the case of I394 HOV vs. general purpose lanes; and (2) draw implications about the potential impacts of the two transportation management alternatives on the statewide transportation network systems.
Now imagine Phase I of the project was just finished and the project team in your firm will meet with MnDOT' s Technical Advisory Panel (TAP) in two weeks to report the major findings at this stage. You are a member of the project team, your supervisor asks you to write up an analysis memo in two week based on the data and findings from Phase I of the study and prepare her for the TAP meeting.
In particular, your memo should include the following six parts:
First, problem statement
Corresponding to two phases of the project, two problems need to be addressed:
(1) At corridor level, estimate current (2000) and future (2020) impacts of I-394 HOV lanes vs. general purpose lanes (addressed in Phase I). In your memo, you should analyze and summarize the data and findings from phase I to address problem one.
(2) Draw implications about potential impacts of the two policy options, as applied to statewide network links (addressed in Phase II).
Problem two will be addressed in the next step of this project. However, in your memo, you need to (a) provide preliminary discussion of the implications based on findings from phase I; and (b) describe the approaches for addressing problem two in Phase II.
Second, methodology
For phase I, major work of the project was to forecast the impacts of opening HOV lanes to the general traffic, in terms of traffic operations and mode split, using a regional travel demand model maintained by the Metropolitan Council. The regional travel demand model is a traditional four-step transportation model developed by the Metropolitan Council and the Minnesota DOT in early 1990s based on Census data and the Twin Cities Travel Behavior Inventory (TBI). In your memo, you should explain the travel demand model and the underline rationale of the model as the project methodology. Additional information about the regional travel demand model can be found from this link:
http://www.dot.state.mn.us/information/hov/pdfs/appendix_b.pdf
Third, data input, output and findings
The data inputs into the model contain (1) node data, depicting characteristics of origins and destinations including population, employment and other factors; and (2) network data describing characteristics of highway and transit networks, such as, travel time on link, average speed, capacity, direction, and transit headway. In this project, the model is updated with both 2000 and 2020 (estimated) data. More information is available on the first six pages of the document under the following link: http://www.dot.state.mn.us/information/hov/pdfs/appendix_e.pdf .
The regional travel demand model outputs the following Measurements of Effectiveness (MOEs):
- Estimated travel speed for HOV
- Estimated vehicle volume at bottleneck locations (AM peak and PM peak)
- Estimated person volume at bottleneck location
- Volume capacity ratio at bottleneck location
- Vehicle Mile Traveled
- Vehicle Hour Traveled
- Travel Time Reliability
- Shift in total trip
- Transit Usage of HOV lane (AM, PM peak period)
These data outputs are presented in the last section of this assignment sheet. You can select information as needed for your analysis.
Fourth, evaluation and analysis
Based on the output MOEs, perform a system analysis and draw implications from two policy alternatives: HOV lanes vs. general purpose lanes. To perform a system analysis, the following parts must be addressed:
(1) Define the system
(2) Generate and assess alternatives available to management (policy scenarios)
(3) Choose alternatives and implementation (run the model)
(4) Compute and summarize MOEs from model output, and
(5) Evaluate and select the preferred policy scenario
Five, discuss implications of HOV vs. general purpose lanes on network system
Briefly discuss (a) the preliminary implications you can draw from Phase I study on the system impacts of the two policy options; and (b) discuss further study, methodology, and data needs for addressing this question thoroughly in Phase II.
Six, conclusions and recommendations
Draw conclusions from your findings. Make recommendations to assist decision making at corridor level. Discuss implications of the two transportation policy options on statewide network system.
Finally, your report should be no longer than 2500 words, 1.5 or double spaced, using 12-point Times New Roman font.
Data output from Phase I (I-394 HOV vs. General Purpose Lanes Impact Study)
Appendix C: Pre-Assignment Survey: Student Information Survey
Student Information Survey
Designed for: CE5214 Transportation Systems Analysis, Spring 2004
Instruction:
- Check all that apply
- Check the version of homework 4 you' re working on:
- For items that are not applicable, check NA
The last four digit of your student ID:
1. Your age: 2. Your gender: female male
3. You are Graduate student 4. Your department:
6. Your main research interests:
7. Years of working experience in the field of transportation planning or engineering: (in year)
8. Do you plan to pursue a transportation career? (Check one) Yes No NA
9. Your weekly computer usage: (in hour)
10. How would you rate your computer proficiency?
11. Of the following words, check any one(s) that describe your learning style or learning preferences:
Reflective observation (watching)
Abstract conceptualization (thinking)
Active experimentation (doing)
12. Between Sensing (concrete, practical facts and procedure oriented), and Intuitive (conceptual, innovative, theories and meanings oriented), which of the following options describe your learning preference best:
13. Between Visual (prefer visual representations, i.e. charts) and Verbal (prefer written or spoken explanations), which of the following options describe your learning preferences best:
Highly Visual
Moderately Visual
Mildly Visual or Verbal
Moderately Verbal
Highly Verbal
14. Between Active (doing) and Reflective (thinking), which of the following options describe your learning preferences best:
Highly Active
Moderately Active
Mildly Active or Reflective
Moderately Reflective
Highly Reflective
15. Between Sequential (linear, orderly, learn in small increments) and Global (holistic, system thinkers, learn in large steps)
Highly Sequential
Moderately Sequential
Mildly Sequential or Global
Moderately Global
Highly Global
16. Innovative teaching strategies motivate you to learn.
17. Rate from 1 to 5 between the two statements: 1: I search for information I' m interested to learn vs. 5: I am comfortable with learning what instructors teach :
1 2 3 4 5 NA
18. List transportation planning or engineering courses you took before:
19. Your familiarity with four-step transportation planning model:
None Very familiar (NA)
20. Your familiarity with simulation in transportation:
None Very familiar (NA)
21. Rate your ability to identify relationships among components of transportation systems:
Poor Very Good (NA)
22. Rate your ability to form opinions regarding transportation issues:
Poor Very Good (NA)
23. Rate your ability to evaluate alternatives by discerning and comparing strengths and weaknesses:
Poor Very Good (NA)
24. Rate your ability to identify information needed to solve a problem:
Poor Very Good (NA)
25. Rate your ability to apply an abstract concept or idea to a real problem or situation:
Poor Very Good (NA)
26. Rate your ability to divide problem into manageable components:
Poor Very Good (NA)
27. Rate your ability to develop several methods which might be used to solve a problem:
Poor Very Good (NA)
28. Rate your ability to use established criteria to evaluate and priotize solutions:
Poor Very Good (NA)
Appendix D: Post-Assignment Survey: Assignment Evaluation Questionnaire
Homework 4 Evaluation Questionnaire
Designed for: CE5214 Transportation Systems Analysis, spring 2004
Instruction:
- Check all that apply
- Check the version of homework 4 you worked on:
Homework 4A
Homework 4B
- For items that are not applicable, check NA
The last four digit of your student ID:
1. Through Homework 4, your understanding on four-step transportation planning model has been improved:
Strongly Disagree Strongly Agree (NA)
2. Through Homework 4, your understanding on transportation simulation has been improved:
Strongly Disagree Strongly Agree (NA)
3. Through Homework 4, your understanding on the development of transportation network patterns has been improved:
Strongly Disagree Strongly Agree (NA)
4. List at least five factors that you think would influence how transportation network develop:
5. Rate your ability to identify relationships among components of transportation systems:
Poor Very Good (NA)
6. Rate your ability to form opinions regarding transportation issues:
Poor Very Good (NA)
7. Rate your ability to evaluate alternatives by discerning and comparing strengths and weaknesses:
Poor Very Good (NA)
8. Rate your ability to identify information needed to solve a problem:
Poor Very Good (NA)
9. Rate your ability to apply an abstract concept or idea to a real problem or situation:
Poor Very Good (NA)
10. Rate your ability to divide problem into manageable components:
Poor Very Good (NA)
11. Rate your ability to develop several methods which might be used to solve a problem:
Poor Very Good (NA)
12. Rate your ability to use established criteria to evaluate and priotize solutions:
Poor Very Good (NA)
13. Rate your overall learning experience with Homework 4:
Poor Great (NA)
14. In homework 4, I had opportunities to practice skills I learned in the course.
Strongly Disagree Strongly Agree (NA)
15. My learning was enhanced through practical experiences ( doing ) provided in the assignment.
Strongly Disagree Strongly Agree (NA)
16. The interactive approach applied in Homework 4 enhanced my learning:
Strongly Disagree Strongly Agree (NA)
17. I was motivated by the teaching strategy applied in Homework 4:
Strongly Disagree Strongly Agree (NA)
18. Any comments regarding your learning experience with Homework 4:
19. Rate your overall satisfaction with Homework 4
Hardly Very Much (NA)
20. Rate your overall satisfaction with SONG 1.0
Hardly Very Much (NA)
21. Rate the quality of SONG 1.0 as a learning tool:
Poor Very High (NA)
22. Rate the quality of the design of Homework 4:
Poor Very High (NA)
23. Rate the quality of TA s instruction:
Poor Very High (NA)
24. Any other comments regarding the design of Homework 4:
25. Learning to perform simulation with SONG 1.0 was easy:
Strongly Disagree Strongly Agree (NA)
26. Supporting information (assignment sheet, instruction, and etc.) for Homework 4 was sufficient:
Strongly Disagree Strongly Agree (NA)
27. Supporting information for Homework 4 was clear:
Strongly Disagree Strongly Agree (NA)
28. Any comments regarding the easiness of performing assignments in Homework 4:
29. Any other comments regarding SONG 1.0:
30. Time you spent on homework 4 (in hours):
31. I m satisfied with the amount of time I spent on Homework 4:
Strongly Disagree Strongly Agree (NA)
32. I can effectively complete Homework 4:
Strongly Disagree Strongly Agree (NA)
33. Any other comments regarding the effectiveness of completing the assignments in Homework 4:
Appendix E: Exam Questions
Exam Questions on Network Growth
CE5214 Transportation Systems Analysis (Spring 2004)
1. Discuss the strengths and weaknesses of four-step planning model. Use examples from at least 2 in-class presentation to illustrate your argument.
[Learning on demand modeling process]
2. What kind of information you would need in order to make decisions on transportation infrastructure investment. What kind of data you can get? How would you evaluate alternatives in the decision-making? Use Examples from at least 2 in-class presentation to illustrate your argument.
[Assess problem-solving skills]
3. Describe the three Twin Cities transportation network systems illustrated in the maps attached, and explain how and why the network patterns are different. What do you think would be the factor(s) that determine the growth and decline of each system? Use examples from at least 2 in-class presentation to illustrate your argument.
[Understand transportation network development process, the influencing factors and players]
|
|
|
Map A Map B Map C