Scholarly article on topic 'An Analysis of Out-of-home Non-work Activity Time Use and Timing Behaviour Based on Work Schedule and Trip Time'

An Analysis of Out-of-home Non-work Activity Time Use and Timing Behaviour Based on Work Schedule and Trip Time Academic research paper on "Economics and business"

CC BY-NC-ND
0
0
Share paper
Academic journal
Procedia Engineering
OECD Field of science
Keywords
{"activity timing" / "time allocation" / "discrete-continuous simulation equations" / "induced and reduce trips"}

Abstract of research paper on Economics and business, author of scientific article — Melawaty Agustien, Ade Sjafruddin, Harun Al Rasyid S. Lubis, Sony S. Wibowo

Abstract This paper attempts to integrate the discrete of activity timing and the continuous data of time allocation based on the concept of utility. The analysis combines the concept of discrete-continuous simulation equations framework to represent the interactions between activity timing and time allocation behaviour during the day. Activity timing choice was modelled as a discrete choice variable involving three alternatives broad periods: home-to-work period, work-based period and post home period. For fixed time workers, it was found that the model in which activity time allocation is assumed to be determined first influence activity timing show statistical measure of fit. Significant parameter of trip characteristics and social economic variables also influence activity timing choice. Activity timing choice model based on time-allocation data can be used to estimate the number of induced and reduce trips and evaluate the impacts of alternative transportation improvement projects to decision to carry out the activities and trips.

Academic research paper on topic "An Analysis of Out-of-home Non-work Activity Time Use and Timing Behaviour Based on Work Schedule and Trip Time"

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Procedía Engineering 125 (2015) 504 - 511

Procedía Engineering

www.elsevier.com/locate/procedia

The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5)

An analysis of out-of-home non-work activity time use and timing behaviour based on work schedule and trip time

Melawaty Agustiena*, Ade Sjafruddina, Harun Al Rasyid S. Lubisa, Sony S.Wibowoa

a Faculty of Civil Engineering , Bandung Institute of Technology, Ganesha 10, Bandung 40132, Indonesia

Abstract

This paper attempts to integrate the discrete of activity timing and the continuous data of time allocation based on the concept of utility. The analysis combines the concept of discrete-continuous simulation equations framework to represent the interactions between activity timing and time allocation behaviour during the day. Activity timing choice was modelled as a discrete choice variable involving three alternatives broad periods: home-to-work period, work-based period and post home period. For fixed time workers, it was found that the model in which activity time allocation is assumed to be determined first influence activity timing show statistical measure of fit. Significant parameter of trip characteristics and social economic variables also influence activity timing choice. Activity timing choice model based on time-allocation data can be used to estimate the number of induced and reduce trips and evaluate the impacts of alternative transportation improvement projects to decision to carry out the activities and trips.

© 2015 The Authors. Published by ElsevierLtd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-reviewunderresponsibility of organizing committee of The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5)

Keywords: activity timing; time allocation; discrete-continuous simulation equations; induced and reduce trips

1. Introduction

The activity based approach analyses people's activities to obtain information about their travel behavior. This allows the analyst to find relationships between the trips and the activities carried out at based on the premise that travel demand is determined by the demand for an activity involving a trip [123]. A relationship exists between activities and trips will influence the choice process of the type of activity, departure time, destination and mode

* Corresponding author. E-mail address: melawaty15@gmail.com;

1877-7058 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of organizing committee of The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5) doi: 10. 1016/j .proeng .2015.11.044

[4,5,6] . Activity engagement can be represented with time use; an individual's activity engagement over a period can be identified by knowing how individual uses time over the period [7,8,9]. Therefore, time use is recognized as a critical subject area for travel behavior analysis and demand forecasting, in particular in activity-based studies.

This study contributes to the literature on activity time-use and activity timing analysis by developing a comprehensive out-of-home non-work activity generation model for workers that considers daily activity time-use behavior and activity timing preferences in a discrete-continue model framework. More specifically, a random utility maximization-based model is formulated to predict workers' activity timing and time allocation patterns in four out-of-home non-work activity purposes at various time periods of the day: (1) Shopping, (2) Sport, (3) Personal business, and (4) Socializing. The time periods of the day are defined based on the representation framework used by [10] to describe the daily activity-travel patterns of workers. According to this framework and the temporal fixities of the work schedule, a worker's day is divided into the following three broad time periods: home-to-work period, work-based period and post home period.

From the viewpoint that activity timing choice is a result of time allocation to activities, a time allocation model is developed in this study to describe activity timing choice. The time allocation model is based on the utilitarian resource allocation theory, that an individual allocates available time to each activity such that the total utility derived from all activities will be maximized. Using this model, activity timing is represented along a continuous time scale. Activity timing choice model is developed as a multinomial logit model with alternatives representing the choice between three broad periods. Time availability constraint for non-work activities would also vary depending on time allocation of work activity before and after the non-work activity and the trip time.

The model in the paper can, therefore, serve as an important component of a comprehensive behavioural tool to analyse the impact of policy actions or changes in household/individual demographics on activity-travel pattern. For example, a policy action that releases some workers at 4 pm instead of 5 pm (as part of either a work staggering policy or an early-release policy to reduce peak-period traffic congestion) may not have the intended effect because such workers may make more out-of-home activity stops after work (either during the commute, or after arriving home at the end of the commute). Even those workers who do not change the number of out-of-home activity stops may now spend more time at each out of home non-work activities. Another possible response of individuals may be to shift non-work stops made earlier during the day to the evening commute and/or the post home-arrival period. From the example we can see that individuals may change their activity-travel behaviour using a combination based on their responses to the policy. These potentially complex responses to policy actions in (a) participation in non-work activities (by activity type), (b) duration of participation, and (c) timing of participation can all be examined using the proposed comprehensive model system.

The additional equation describes another problem of discrete-continuous resource allocation that can be represented by a modified multinomial logit model. This model estimates the variations in trade-off between trip time and time allocated to activities in relation to the exogenous variables characterizing the transportation system. The objective of this work is therefore to segregate the time allocated to trips from the total time dedicated to out-of-home activity participation so as to provide a tool capable of capturing the direct consequences on time allocation to non-work activity trips of modifications to an individual's time budget. For example, an improvement in the level of service of a transport system will reduce trip time and time thus saved can be spent in activities that do not generate trips or activities that do.

The next part of this paper is organized as follows. Section 2 explained the discrete-continuous simultaneous equations system. Description of socio-economic, activities and trips respondents data are described in Section 3. Section 4 explains the procedures and the results of the estimated parameter values of the model. Finally, Section 5 outlines the conclusions of this study.

2. Concept Of Discrete-Continuous Simulation Equations Framework To Represent The Interactions Between Activity Timing and Time Use

A discrete-continuous choice model of time allocation to two or more types of non-work activities is formulated based on the concept of random utility maximization. The model is developed by [9] to explore the relationship between activity timing and duration of maintenance activities. The model development takes into account the discreteness inherent in daily activities engagement and time allocation (i.e. one type of activity may not be engaged,

therefore no time may be allocated to it at all, during a day). Without requiring simplifying assumptions other than the one that at least one type of the two types of activities will be engaged, the utility model is formulated as discrete and continuous equation systems. A discrete or continuous equation systems, one can provide a construct to link the discrete and continuous components. The most obvious of these constructs is a reduced form approach. A common way to implement a reduced is to start with discrete model.

Let Ti„ be a linier function that determines discrete outcome i for observation n and let yi„ be the corresponding continuous variable in the discrete/continuous modelling system. The discrete and continuous equation systems can be written as follows:

Tin = PiX in +ky in +Sin (1)

: a vector of estimable parameters for discrete outcome i Xi„ : a vector of observable characteristic (covariates) that determines discrete outcomes for observation „ s'„ : a disturbance form

^ : estimable parameters

The corresponding continuous equation be the linear function:

y. =aw. + v.

i : a vector of estimable parameters for the continuous variable observed for discrete outcome i Wi„: a vector of observable characteristic (covariates) that determines yi„ vi„ : a disturbance form

Equation (2) is estimated using ordinary least square with appropriate selectivity bias correction (such as adding a selectivity bias correction term). For estimation of the discrete outcome portion of the discrete/continuous process, note that yi„ is endogenous in equation (1) because yi„ changes with changing T„ due to interrelated discrete/continuous structure. To avoid estimation problems, a reduced form model is obtained by substituting equation (2) into equation (1), giving

T„ =fiiX„ +0i (dlWm + v„) + (3)

With equation (3) a discrete outcome model is derived readily. For example if the g are assumed to be generalized

extreme value distributed, a multinomial logit model results with the probability of observations „ having outcome i (P„(i))as:

exp(PtX„ +h6Wi„)

P„ (i) =

£ exp(PiXi„ +№Win)

vi (4)

3. Data Description

Palembang is the capital city of Indonesia's Sumatera Selatan (South Sumatera) Province. The majority of the city is located north of the river, where the old City Centre is also located. The area to the south of the river is less developed, although a second commercial centre is being promoted, the city's parliament is already located there, and there are plans to relocate provincial government offices. The current population of Palembang city is estimated to be about 1.8 million [3]. In recent years this population has been growing steadily at around 2% per annum. If this trend continues, the population in 2020 will be around 1.98 million and in 2030 will be around 2.40 million. The city centre itself is not so densely populated, largely because of the presence of offices, shops and other commercial and government buildings.

The characteristics of an activity that was observed in this study consisted of the activity type and its time allocation. The type of non-work activities that are observed in this study is out of home non-work activities. Non-work activities can be done inside home or outside home in which the timing, time allocation and location are more flexible than the mandatory and maintenance activities. There are respondents who have more than one non-work activity in a day and

16% respondents choose shopping as their discretionary activity. Percentage based on activity type and average time allocated to different non-work activities are described on Figure 1 and Figure 2.

Fig. 1. Percentage of Respondent Based on Non-Work Activity Type

Fig. 2. Percentage of Respondent Based on Average Time Allocated to Different Non-Work Activities

The data set was obtained from an individual trip diary survey that was conducted in 2014 in Palembang City, Indonesia. The information from the data collection consists of socio-economic condition of household and individual, demographic characteristics along with detailed information about trips undertaken over a 24 hour period. Non-work activities were observed for 300 respondents where there were 420 non-work activities that were conducted in a day by these respondents. Table 1 below illustrate the composition of the sample, individual and household characteristics. The average household size for the sample of 300 respondents is 3 to 6 people per household. Average vehicle ownership is 2 units per household with 40% of respondents having 2 units cars and 27% of respondents having one or more motorcycle, the rest have cars and motorcycles. Average household monthly income of the respondents is IDR 10,000,001 - IDR. 12,500,000.

Figure 3 and 4 explain the cross tabulation in which the distribution of time allocation and activities timing are related to each other. The pictures provide us with a better understanding the relationship between non-work activities timing and time allocation. The decision to perform non-work activities based on a combination of time periods can be divided into 3 time interval, namely: the period of time before work that is on the way from home to the office (home-to-work period), the period of lunch break time during working hours (work-based period) and the period after returning home (post-home period). The decision to perform the activity is influenced by the allocation of time required to maximize the utility derived from such activities. Based on the survey result, there are three time intervals for doing discretionary activities. The intervals consist of time allocation less than 1 hour, between 1 -3 hours and more than 3 hours.

s e R f o

Non-Work Activity Time Allocation Fig. 3. Percentage of Activity Timing Based on Time Allocation of Non-Work Activities

100% 80% 60% 40% 20% 0%

■ Home-to-work period

■ Work-based period

, TT , ,TT , Post-home period

< 1 Hour 1-3Hours > 3 1

100% 80% 60% 40% 20% 0%

iJ h- J

Work-Home Work Based Post Home Non-Work Activity Timing

l< 1 Hour ■ 1-3Hours I > 3 Hours

Fig. 4. Percentage of Time Allocation Based on Activity Timing of Non-Work Activities Table 1. Individual and Household Social Economy Characteristics

Characteristics Clasification Number of Respondents (Person) Percentage (%)

Age Less than 26 years Person age 26-40 years Person age 41-55 years More than 56 years 15 181 84 20 5 60 28 7

Type of Work Government employee Private employee 177 123 59 41

Average household size < 3 person household 3 - 6 person household > 6 person household 69 117 114 23 39 38

Average car ownership 0 1 unit 36 87 12 29

Characteristics Clasification Number of Respondents (Person) Percentage (%)

Average car ownership 2 units 120 40

More than 3 units 57 19

Average motorcycle ownership 0 1 unit 2 units More than 3 units 60 93 111 36 20 31 37 12

Average household monthly income

IDR 5,000,001 - IDR 7,500,000 IDR 7,500,001 - IDR 10,000,000 IDR 10,000,001 - IDR 12,500,000 IDR 12,500,001 - IDR 15,000,000 More than IDR 15,000,000

75 54 102 42 27

25 18 34 14 9

From the combination between alternative time allocation and activity timing for conducting out of home non-work activities, it is appear that respondents who allocate their time to do the out of home non-work activities less than 1 hour and 1-3 hours tend to do the activities at home-work and work-based period. Whereas respondents who allocate their time more than 3 hours tend to do the activities at post-home period. Activity timing choice influenced by the type of activity, time allocation of more than 3 hours is generally used for entertainment activities which is need more time to do the activities. Routine activity such as pick up children, lunch and going to the market have less time allocation and conduct at before work and work based period.

4. Empirical Result

The decisions regarding the time of day of activity participation and activity time-allocation are modelled using a joint discrete/continuous econometric framework. In such joint systems, logical consistency considerations require certain restrictions to be maintained on the coefficients representing the causal effects of the dependent variables on one another. Specifically, in the context of the joint time of day of activity participation and activity time allocation model of the current paper, the restrictions imply a recursive causal model in which activity duration affects the activity participation decision. Time allocation is a function of socio-economic characteristics, type of activity, and the travel characteristics associated with the activities that will be conducted. After a specified time allocation, the decision to perform the activity is determined by a function of the time allocation of activity, activity type and the socio-economic

condition. Based on these models, it is possible that a different model structure is suitable for different segments of the respondents, different types of activities and different urban areas.

This section will explain the estimation results model for respondents with a fixed working time (fixed-time workers). The first block corresponds to the activity participation choice model, where work-based activity timing is considered as the base alternative. The second block shows the log-linear duration model. The standard deviation of the error term in the log-linear duration model is also provided.

Table 2 provides the model estimation results for multinomial logit choice for 3 alternatives of activity timing. The first row on the table is considered the base alternative. Various socio-demographic activity variables are found to influence activity timing choice. The alternative specific constant in the model show that, relative to alternative work-based period, there is a greater likelihood of pursuing a non-work activity in the home-work and post-home period. All of the alternatives specific constants positive and significant. The greatest likelihood occurs in the variable posthome period, and it shows that the propensity to pursue the out of home non-work activities in the after returning home period or post-home period is quite plausible, compared with other alternatives.

Table 2. Discretionary Activity Timing Model Based on Time Allocation

Variable Parameter Estimate t-statistic

Alternative:Work Based Model = 0

Home-Work Period

Constant 2.7741 8.6421

Person age 26-40 years 1.9427 2.1289

Public Servants -2.6713 -7.9537

Household income -2.1431 -5.8172

Trip time from office -2.1431 -25.8172

Joint tour 2.5499 6.7631

Post- Home Period

Constant 4,2013 25.2441

Activity duration 1.5113 12.3271

Household income 2.0872 13.7631

Private Employment 3.6522 23.8611

Car ownership -2.2463 -3.7621

Trip time from home -2.0463 -12.3441

Joint tour 2.5422 12.9632

Log-Linear Duration Model

Constant 1.8907 25.633

Public Servants* -0.058 -1.217

Private Employment -0.191 -3.979

Entrepreneurs* 0.061 1.278

Number of school age family member -0.159 -6.266

Household income 0.199 6.033

Car ownership 0.156 3.507

Motorcycle ownership* -0.058 -1.224

Log-Linear Model Error Term

Standard deviation, o 0.7385 43.744

Error Correlation Estimates

T midday-duration -0.5984 -16.895

*Not significant at the a=0.1

The modelling shows that there were some significant and not significant variables are explained in table 2. Private employment appear to show a greater propensity to pursue out of home non-work activities in the after returning home period and public servant tend not to pursue out of home non work activities in the home-work period.This may be due to the tight of work schedule and the need to allocate more time to do out of home non work activities. Another significant variable is person age 26-40 years and joint tour which have positive coefficient value. The workers aged 26-40 years appear to show a greater propensity to pursue non-work activities in the home-work period. This may be a serve-child trip or another habitual trip that the worker performs before work schedule. The activity duration variable affects activity-timing choice in post-home periode. The coefficient associated with the duration variable is positive indicating that worker is not constrained with respect to the length of their activity in these-periods. They can pursue

long activity time in the post-home period (figure 3 also show this). The positive sign on household income show that worker who have greater income tend to do non work activities in post-home period while the negative sign on number car ownership show that respondents who have greater number of car ownership do the non-work activities in period apart from post home period,

Trip characteristics also influence the choice of activity timing. The workers who have a shorter distance from office to a non-work activity location tend to do the activity in home-work period. Otherwise, the workers who have a shorter distance from home to a non-work activity location tend to do the activity in post-home period. Number of people who joint the tour for doing non work activity is also influence activity timing and it is occurred in the home work and post home period.

The log duration model also effect plausible indications. Workers with type of work public servants and private employment indicating the ability to enggage in shorter non-work activities time allocation relative to workers with type of work entrepreneurs. In this model, the error correlation between post home period activity participation and activity duration is the only statistically significant error correlation. The coefficient suggest that there is a positive correlation between post home participation and activity duration during this period. It indicates that unobserved factors that increase the likelihood of participation during the post home period also increase activity duration during that period.

5. Conslusions

This paper has presented an exploration of the relationship between activity timing and activity time allocation for non-work activities such as shopping, social, sport, and others. The study involved the analysis of joint models of activity timing and duration for fixed time workers while allowing error correlations between the activity timing and time allocation model equations. Activity participation choice was modelled as a discrete choice variable involving three alternatives broad periods: home-to-work period, work-based period and post home period.

The modeling system in this study primarily adopts a modified version of the frameworks proposed by [12] to explore the relationship between activity timing and duration of maintenance activities. The data set was obtained from an individual activities and trips diary survey that was conducted in 2014 in Palembang City, Indonesia. The information from the data collection consists of socio-economic condition of household and individual, demographic characteristics along with detailed information about activities and trips undertaken over a 24 hour period.

For fixed time workers sample, it was found that the model in which activity duration is assumed to be determined first influence activity timing show statistical measure of fit. Significant parameter of activity duration variable show that utility value of out of home non-work activities which is represented by time allocation influence activity timing choice. While the positive sign of parameter of activity duration variable show that if the respondents need more time to do the non-work activities so they prefer to choose post-home period activity timing compared with other alternatives.

Variation of value of parameters and constant in the model show that there is a time period alternative that is tend to choose based on respondents social economics background and characteristics of the non-work activities and trips. Furthermore, these model show how travel demand modelling systems based on time-allocation data can be used to estimate the number of induced and reduce trips that would result from travel time, and evaluate the impacts of alternative transportation improvement projects to decision to carry out the activities and trips.

References

[1] Fujiwara, et.al.,: Modelling the Interaction between Activity Participation and Time Use Behaviour over the Course of a Day, Journal EASTS, Vol.8 (2010)

[2] Meloni, I., E. Spissu, and M. Bez, A Model of the Dynamic Process of Time Allocation to Discretionary Activities, Journal of Transportation Science, Vol. 41, No. 1, (2007). pp. 15-28.

[3] Shiftan Yoram, Ben-Akiva, M.E. , Activity Based Modeling As A Tool For Better Understanding Travel Behaviour, 10th Intrnational Confrnc of Travl Bhaviour Rsarch, Lucrn 10-15 August, (2003).

[4] Bhat, C.R., and Koppelman, Activity Based Modeling of Travel Demand. Transportation Research, Vol. 15, No. 1-2, (2005a).

[5] Jones, P.M., Dix, M.C., Clarke, M.I. and Heggie, I.G. Understanding travel behaviour,Aldershot: Gower. (1983),

[6] Kittamura R., and Satoshi Fuji "Time-use data, analysis and modeling: toward the next generation of transportation planning methodologies". Tra„sportpolicy, Vol. 4, No. 4, , pp. 225-235. Elsevier Science Ltd, (1997).

[7] Pas, E.I. and Harvey, A.S. "Time use research and travel demand analysis and modeling", In Stopher, P.R. and Lee-Gosselin, M Understanding Travel Behavior in an Era of Change, Elsevier, Oxford 315-338. (1997).

[8] Srinivasan, K.K., and S.R. Athuru., Analysis of Within-Household Effects and Between Household Differences in Maintenance Activity Allocation, Transportation, Vol. 32, No. 5, pp. 495-521, (2005).

[9] Yamamoto, T., dan R. Kitamura, An Analysis of Time Allocation to In-Home and Out-of-Home Discretionary Activities across Working Days and Non-Working Days, Transportation, Vol. 26, No. 2, (1999), pp. 211-230.

[10] Bhat, C.R , et.al, Modeling Household Interactions in Daily In-Home and Out-of-Home Maintenance Activity Participation, Transportation Research Record, (2005b).

[11] Badan Pusat Statistik, (2014): Palembang Dalam Angka

[12] Pendyala, R.M. and Bhat, C.R. An Exploration Of The Relationship Between Timing And Duration Of Maintenance Activities, Journal of Transportation Science, Vol. 81, No. 1, (2010). pp. 35-48.,