Content uploaded by Ji-Young Song
Author content
All content in this area was uploaded by Ji-Young Song on Jun 01, 2021
Content may be subject to copyright.
PLEASE SCROLL DOWN FOR ARTICLE
This article was downloaded by:
[Lorch, Richard]
On:
7 May 2009
Access details:
Access Details: [subscription number 790465784]
Publisher
Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
Building Research & Information
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713694730
Modelling the use of space and time in the knowledge economy
William Fawcett a; Ji-Young Song a
a The Martin Centre for Architectural and Urban Studies, Cambridge University Department of Architecture,
Cambridge, UK
Online Publication Date: 01 May 2009
To cite this Article Fawcett, William and Song, Ji-Young(2009)'Modelling the use of space and time in the knowledge
economy',Building Research & Information,37:3,312 — 324
To link to this Article: DOI: 10.1080/09613210902863682
URL: http://dx.doi.org/10.1080/09613210902863682
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
INFORMATION PAPER
Modelling the use of space and time
in the knowledge economy
William Fawcett and Ji-Young Song
The Martin Centre for Architectural and Urban Studies,Cambridge University Department of Architecture,
1 Scroope Terrace,Cambridge, CB21PX, UK
E-mails: wf223@cam.ac.uk and jys22@cam.ac.uk
Flexible working gives employees in knowledge-based organizations new opportunities for choosing the locations and
times of work activities. This trend has been described many times, but the literature has little quantified data about
the resulting activity patterns, or their impact on the scale of demand in buildings. A preliminary simulation model of
individual employees’ decision-making in office-based organizations was developed, generating quantified output data
describing the times and places chosen for work activities. Decision-making was based on individual preferences
between home and office locations over a 25-time period weekly cycle. Systematic models runs provided indications
of possible trends. Survey data from real organizations that compared participants’ actual and preferred activity
patterns provided some empirical support for the model findings. The model requires further empirical validation,
and offers scope for enhancement. Information provided by models of this type would be highly relevant for the
briefing, design, and management of buildings for the knowledge economy.
Keywords: agent-based simulation, facilities management, flexible working, home working, knowledge economy, time
use, utilization, work– life integration
Les horaires de travail variables donnent aux employe
´s des organisations fonde
´es sur le savoir de nouvelles possibilite
´sde
choisir les lieux d’exercice et les horaires de leurs activite
´s professionnelles. Cette tendance a e
´te
´de
´crite a
`de nombreuses
reprises, mais la documentation existante comporte peu de donne
´es quantifie
´es sur les sche
´mas d’activite
´qui en re
´sultent,
ou sur leur impact sur l’e
´chelle des demandes dans les immeubles. Un mode
`le de simulation pre
´liminaire de la prise de
de
´cision par chaque employe
´dans des organisations fonde
´es sur le travail de bureau a e
´te
´de
´veloppe
´, ce qui a ge
´ne
´re
´des
donne
´es de sortie quantifie
´es de
´crivant les horaires et les lieux choisis pour les activite
´s professionnelles. La prise de
de
´cision e
´tait base
´e sur les pre
´fe
´rences individuelles entre travail a
`domicile et travail au bureau sur un cycle
hebdomadaire de 25 tranches horaires. L’exe
´cution syste
´matique des mode
`les a fourni des indications sur les
tendances possibles. Les donne
´es d’enque
ˆtes re
´alise
´es aupre
`s d’organisations re
´elles, comparant les sche
´mas d’activite
´
re
´els et pre
´fe
´re
´s des participants, ont assure
´une certaine corroboration empirique des re
´sultats du mode
`le. Ce mode
`le
ne
´cessite une validation empirique plus approfondie et offre des possibilite
´s d’ame
´lioration. Les informations fournies
par les mode
`les de ce type seraient hautement pertinentes pour le cahier des charges, la conception et la gestion
d’immeubles destine
´sa
`l’e
´conomie du savoir.
Mots cle´s: simulation base
´e sur les agents, gestion globale de ba
ˆtiments, horaires de travail variables, travail a
`domicile,
e
´conomie du savoir, emploi du temps, utilisation, inte
´gration travail-vie personnelle
Introduction
It is widely accepted that working patterns are chan-
ging in the emerging ‘knowledge economy’ of devel-
oped countries. Crucial drivers are as follows:
.mobile telecoms and distributed computing
.educated, self-motivating and highly valued
employees
This results in a considerable expansion of individual
choice about how, when and where work activities
take place. Work activities by employees used to be
BUILDING RESEARCH &INFORMATION (2009) 37 (3), 312–324
Building Research & Information ISSN 0961-3218 print ⁄ISSN 1466-4321 online #2009 Taylor & Francis
http: ⁄ ⁄www.informaworld.com ⁄journals
DOI: 10.1080/09613210902863682
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
concentrated in the employers’ premises during speci-
fied working hours, but are becoming dispersed:
many people can now work in their employer’s pre-
mises, at home, at client sites, in fact more or less any-
where, at any time of the day or night. Commentators
have been discussing these changes for some time; for
example, Duffy (1992) wrote:
The key to the new office interior is the freedom
in use of time which information technology
beings. The nine-to-five office day is anachronis-
tic. The office is likely to become a meeting place
rather than a place for so many desks.
(p. 235)
The study reported in this paper was developed in the
context of office-based organizations, but the principles
should be relevant to any building type in which individ-
ual users have freedom of choice about when and where
they carry out activities; this could apply, for example, to
buildings for higher education and retailing.
There is now an extensive literature on new ways of
working, including:
.surveys of current practices, emphasizing the social
science perspective (for example, Felstead et al.,
2005; Lewis and Cooper, 2005; Strelitz and
Edwards, 2006; and Hyman et al., 2005)
.architectural case studies of new office design (for
example, Myerson and Ross, 2006)
.predictions/speculations about future trends (for
example, Harrison et al., 2003; Worthington,
2005; and DEGW, 2008).
Transportation/telecommuting impacts of new work-
styles are another fertile branch of study (for
example, Robert and Borjesson, 2006).
Most references to changing ways that employees
manage their work activities are qualitative, such as,
‘there are many ways in which to personalize time’
(Felstead et al., 2005, p. 28); ‘The potential of distrib-
uted working in enabling office workers to shape their
work schedules and thus their work/life balance’
(Katsikakis, cited in Worthington, 2005, p. 100); and
‘work teams take control over their use of time’
(Lewis and Cooper, 2005, p. 39). Quantified studies
usually focus on the total hours of work (for
example, Bonney, 2005); the literature contains little
quantified data about day-to-day time management.
There is a consensus that workstyles are changing,
but little information about the quantified impact
of change, creating uncertainty for those involved
in the briefing, design and management of buildings
for organizations in the knowledge economy. The
preliminary study reported here addressed this
problem by investigating possible activity patterns
using an agent-based simulation model.
New workstyles
Employees in the knowledge economy are able to adjust
their activities to fit in with their personal priorities,
not just their employer’s. This is often described as a
move to ‘work–life balance’, ‘work– life integration’ or
‘work–life harmonization’ (Lewis and Cooper, 2005,
p. 8). Due to the great diversity of different people’s non-
work commitments, individual control can be expected
to increase variability between employees’ work patterns.
Both the provision by employers of opportunities for
work– life harmonization and the take-up by employees
are increasing. A 2006 survey in the UK, for example,
reported that about 90% of employees had access to
some kind of flexible working arrangement (Hooker
et al., 2007). In the UK the trend is reinforced by
public policy, with legislation introduced in 2003
giving specified types of employees the right to request
flexible working, and requiring employers to consider
their requests seriously. In 2007 the legislation was
extended to cover a broader range of employees
(Department of Business Enterprise and Regulatory
Reform (DBERR), 2007), and further extensions are
planned. Four-fifths of statutory requests for flexible
work have been accepted by employers, either fully or
partially (Fitzner and Grainger, 2007).
It is proposed that the activity patterns adopted by the
large and increasing number of employees with choice
about the times and places for carrying out work
activities result from individual decision-making in
response to individual constraints and opportunities.
To model this complexity a simulation model was devel-
oped with an element of randomness in the decision-
making of individual employees. The model allows
numerical values to be assigned to some of the factors
that characterize new working practices, and generates
a quantified picture of the resulting pattern of work
activities. An advantage of modelling compared with
case study investigations is that it is not limited to the
study of present-day situations, but allows the explora-
tion of hypothetical scenarios of change.
Conceptual model of employee
decision-making
When they can choose between alternative places and
times for work and non-work activities, we assume
that employees have a decision-making process in
which each alternative is evaluated on the basis of
constraints and personal values, and ranked; and
then the most favourable alternative is selected.
In constructing a model of this decision-making
process we first consider the alternative places that
Modelling the use of space and time in the knowledge economy
313
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
have to be evaluated. We propose that evaluation takes
account of two attributes:
.performance: how good each place is for perform-
ing work tasks
.convenience: how convenient each place is for
dealing with non-work commitments
Thus, each alternative place has a performance score
and a convenience score. The two attributes are inde-
pendent: for example, in the traditional style of office
and home, the office would be given a high score for
work-related performance and a low score for non-
work convenience, whereas the home would be given
a low score for performance and a high score for con-
venience. In more modern environments, the conven-
ience scores of offices would be higher as greater
provision is made for employees’ non-work commit-
ments; and the performance scores of homes would
also be higher with distributed computing.
How do individuals use the two scores to rank
alternatives? We propose thatthey use weighted averages
of the performance and convenience scores, and select
the place with the highest weighted average score. The
relative weighting is specified in a work–life index.
This index varies over time, in contrast to the perform-
ance and convenience scores that are fixed: the varying
work– life index causes different places to be selected at
different times.
Development of the work– life index starts with the
observation that there are strong social conventions
about the hours of the day that are devoted to work
or to personal life, as shown in time-use surveys
(Michelson, 2005). There are strong concentrations of
work activities in weekday mornings and afternoons
(pp. 117–119). This corresponds to a work – life
index in which performance would have high weighting
during weekday mornings and afternoons, and conven-
ience would have high weighting at other times.
In assessing how this conventional time-use pattern
might change in the knowledge economy, it is interest-
ing to note that a comparison between Canadians
working at employers’ workplaces and those working
at home showed only a slight change in the pattern
of time-use (Michelson and Crouse, 2004). A study
of home-based white-collar workers reported that:
None of the respondents described any difficul-
ties with time boundaries at the end of the day
or with switching off from work.
(Halford, 2005, p. 27)
A further survey of ‘nomadic’ workers who relied on
mobile telecoms and distributed computing reported
that they:
expressed a desire to separate business and per-
sonal roles using traditional boundaries of
space and time
(Cousins and Robey, 2005, p. 176)
but despite this desire:
the technology played a key role in extending
business into spaces and times traditionally
reserved for personal activities
(p. 177)
In the knowledge economy, therefore, there does not
seem to be any reason to expect that employees will
move to a radically new time use convention with the
concentration of work activities at other times than
weekday mornings and afternoons. Although some
individuals may choose to, for example, work at night
and use the day for non-work activities, it seems
likely that most people will choose relatively small devi-
ations from the conventional time use pattern.
The core concepts of (1) independent performance and
convenience scores and (2) a time-varying work– life
index form the basis for the computational model of
employee decision-making.
Computational model of employee
decision-making
Performance and convenience scores
Each employee ihas to choose between three alterna-
tives j:
.work in the employer’s office
.work at home or away from the office
.non-work
There could be more alternatives, but the current
model uses these three.
Employee iassigns two scores to each alternative j– one
for performance, p
ij
, a work-centred perspective, and
one for convenience, c
ij
, a non-work-centred perspec-
tive. The performance score takes account of factors
relating to productivity, such as access to data and tele-
coms, and also cultural factors, like the need to be
visible in the office to maintain credibility with senior
staff, or the need for meetings or informal contacts.
This scoring method was used to describe three differ-
ent environments: ‘traditional’, ‘intermediate’ and
‘modern’. The main differences are:
.traditional offices make almost no provision for
non-work commitments so have a very low
Fawcett and Song
314
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
convenience score; this rises in intermediate
and modern offices
.traditional homes make almost no provision for
work activities so have a very low performance
score; this rises in intermediate and modern
homes with distributed computing
Therefore, the convenience scores of offices and the
performance scores of homes both rise in intermediate
and modern environments.
The performance and convenience scores that represent
these environments are shown in Table 1. They are on a
scale from zero (bad) to five (good). The values used for
the preliminary studies were estimates chosen to rep-
resent discrete steps in the move to the knowledge
economy; they were not derived from surveys.
Work^ life index
The model considers five time periods in a day – dawn,
morning, afternoon, evening, and night – over five
working days, making 25 time periods in a week.
Each employee ihas a work– life index, W
i
, which
comprises 25 values, one for each of the 25 time
periods, W
i
¼fw
it
,t¼1, 2, ...,25g. The work –life
index value for each time period is between zero and
one and expresses the relative weight of performance
(work) compared with convenience (non-work) –
high index values mean that work is given priority.
The estimation of values for the work –life index began
with the observed data about the time use of employed
people in Canada in 1998, reported by Michelson
(2005) (Table 2). These data recorded the proportion
of non-home-based employees who were at home at
different times of the day, and this proportion was
used as a measure of the priority given to non-work
activities; time away from home was used as a
measure of the priority given to work activities.
However, when away from home people would
have carried out other activities as well as working,
so the reference work–life index may overestimate
the weight given to work activities. If better empirical
data can be obtained, the values could be substituted
in the simulation model.
The source data were an average over all weekdays.
These values were adjusted for the different days
of the week in accordance with travel to work data
for the UK (Department for Transport (DfT), 2005),
which suggested that there were slightly higher than
average work activity on Wednesdays and Thursdays,
and lower on Fridays, although the differences were
quite small. The average weekday values were multi-
plied by adjustment factors derived from the travel
data (Table 3). This gave a set of average, or ‘reference’
work–life index values for the 25 time periods in a
week (Table 4).
In the simulation model each individual employee’s
work–life index is generated by random deviations
from the reference work–life index, the size of devi-
ations following a normal distribution in which small
deviations occur more often than wide deviations.
The generation is performed with the Excel function
NORMINV which gives the inverse of a cumulative
normal distribution. For employee iin time period t,
Ta b l e 1 Performance and convenience scores of activity ^space alternatives for t raditional, intermediate, and modern environments as
used in systematic model runs
Traditional Intermediate Modern
Performance Convenience Performance Convenience Performance Convenience
O⁄ce304152
Home work 1 2 3 2.75 4.50 3.75
Non-work050505
Ta b l e 2 Conventional daily time-use data
Dawn Morning Afternoon Evening Night
Percentage of employed people away from home 17 79 85 35 10
Percentage of employed people at home 84 21 15 65 90
Note: Values are from percentages of non -home-based workers at home on a working weekday by minutes through the day for Canada,1998.Sourc e:
Michelson (2005), p.60.
Modelling the use of space and time in the knowledge economy
315
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
the work–life index value, w
it
, is given by:
wit ¼NORMINV ðRANDðÞ;rt;
s
Þ
where: RAND( ) is a random number between zero and
one; r
t
is the reference work– life index value for time
period t(the mean of the normal distribution);
s
(sigma) is a parameter controlling the amount of vari-
ation from the mean (the standard deviation of the
normal distribution); values of w
it
are closer to r
t
for
low values of
s
.
This sometimes generates work – life index values that
are greater than one or less than zero. These values
are replaced by random numbers between 0.9 and
1.0 or between zero and 0.1, respectively.
Employee types
The amount of variation from the reference work – life
index is controlled by the parameter
s
(sigma); higher
values of
s
(sigma) generate more variation. In the
model, this parameter was used to distinguish
between two types of employee, ‘unreformed’ and
‘flexible’. The work– life index values of unreformed
employees were assumed to show only small variation
from the reference work– life index, whereas there was
much greater variation for flexible employees.
The work–life index values of unreformed employees
were generated using a value of
s
(sigma) equal to
0.02, and for flexible employees the value of
s
(sigma)
was 0.35. Table 4 shows the reference work – life
index and specimen values for unreformed and flexible
employees. Figure 1 plots the work –life index values for
five unreformed employees, and Figure 2 shows the
work–life index values for five flexible employees.
Note that the flexible employees show much greater
individual variation, but they still inherit the reference
index’s morning and afternoon concentration of work
activities. Because employees’ work– life indexes are
generated as random variations from the reference
work–life index, the average of many employees’
indexes will approximate to the reference index.
Decision-making
The employees’ decision-making procedure for choos-
ing between the ‘office’, ‘home work’ and ‘non-work’
alternatives uses the performance and convenience
scores and the work– life index. A weighted average
attractiveness score a
ijt
is calculated for employee i,
for each alternative j, and each time period t, using
the employee’s work – life index, w
it
:
aijt ¼pij wit þcij ð1witÞ
It is assumed that each employee has working time
budget of ten time periods in a week, so the higher
scoring of the ‘office’ and ‘home work’ alternatives is
identified for all 25 time periods, and then the ten
time periods with the highest of these scores, whether
‘office’ and ‘home work’, are selected. The other 15
time periods are assigned to ‘non-work’.
This completes the simulated week’s diary for one
employee; the same process is carried out for each
employee in turn.
Simulation model
The simulation is implemented in Microsoft Excel.
Each time the model is run it generates the ‘diaries’
of 100 employees over one week.
The input data describing a scenario for a model run
comprise the following:
.environment: performance and convenience scores
for the three activity–space alternatives (office
work, home work, non-work)
.behaviour: the value of
s
(sigma) that generates
each employee’s work– life index values
In a model run it is assumed that all employees share
the same performance and convenience scores, and
that their work– life indexes are all generated with the
same value of
s
(sigma). The model could be adapted
to allow for model input describing more diverse
populations.
Having simulated the diaries of all employees,
the model outputs 100 lines of data specifying the
activity–space alternatives selected by each employee
for the 25 time periods. This data can be analysed in
various ways after the model run.
Ta b l e 3 Daily adjustment to the reference work ^life index
Monday Tuesday Wednesday Thursday Friday
Adjustment factor 1.00 1.00 1.01 1.02 0.97
Notes: Average values in the reference work^ lifeindex were multiplied by these factors to give daily work ^life index values.
Valuesare from focus on personal travel for the UK,1998^ 2003. Source: Department forTransport (DfT) (2005).
Fawcett and Song
316
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
Ta b l e 4 Reference work^life index and typical work^ life index values for unreformed and£exible employees
Monday Tuesday Wednesday Thursday Friday
dma e n dma e ndma e n dm a e ndma e n
Reference work^
life index
0.17 0.79 0.85 0.35 0.10 0.17 0.79 0.85 0.35 0.10 0.17 0.80 0.86 0.36 0.10 0.18 0.80 0.86 0.3 6 0.11 0.15 0.78 0.84 0.3 4 0.08
Unreformed
employee
s
(sigma) ¼0.02
0.16 0.76 0.85 0.36 0.11 0.18 0.77 0.81 0. 33 0.12 0.15 0.79 0.85 0.37 0.0 6 0.16 0.7 7 0.84 0.34 0.09 0.12 0.74 0.82 0.3 0 0.0 8
Flexible employee
Note: d, dawn; m, morning; a, afternoon; e, evening; and n, night.
Modelling the use of space and time in the knowledge economy
317
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
Experiments with simulation model
The simulation model was run for a series of scenarios
based on systematic variation of environment and
behaviour input data. As described, there were three
different environments, ‘traditional’, ‘intermediate’
and ‘modern’, defined by the performance and conven-
ience scores (Table 1); and two types of employee
behaviour, ‘unreformed’ and ‘flexible’, with work –
life index values generated by different values of
s
(sigma) (Table 4). The simulation model was run for
100 employees, for all six combinations of environ-
ment and behaviour (Table 5).
The output of each scenario was presented graphically
to show the number of employees at the employer’s
premises and the number working at home for each
of the 25 time periods in a week; and also the
average and peak demand at the employer’s premises
at each time of the five time periods in a day – this
was averaged over the five days.
The results for unreformed employees are shown in
Figure 3 for traditional, intermediate, and modern
environments. As might be expected, the unreformed
behaviour/traditional environment combination
(U-T) leads to a typical pre-knowledge economy
pattern, with all work being carried out in the employ-
er’s office on weekday mornings and afternoons, and
no home-based work. The move to the intermediate
environment (U-I) brings virtually no change, and in
the modern environment (U-M) there is only a small
amount of home-based work in the morning time
periods. There is no work outside the morning and
afternoon time periods either at the employer’s office
or at home. The simulation suggests that if employees
are strongly rooted in traditional working patterns,
Figure2 Five typical £exible employees work^ life index values:
s
(sig ma) ¼0.35
Figure1 Five typical unreformed employees work^ life index values:
s
(sigma) ¼0.02
Fawcett and Song
318
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
changing the environment at both office and home has
little impact on their decision-making.
The results for flexible employees are shown in Figure 4.
Faced with a traditional environment (F-T), they transfer
about 30% of work from employer’s office at peak times,
some to home-based work at peak times, but mostly to
office- and home-based work at other times – evening,
night and dawn. Approximately 10– 20% of employees
are in the office at non-peak times.
In the intermediate environment (F-I), there is much
the same amount of work in the employer’s office at
peak times, and some increase in office-based work at
non-peak times, but there is a considerable change in
home-based work: the favoured times change form
non-peak to peak times. Home-based work at non-
peak times is greatly reduced.
In the modern environment (F-M), the amount
of office-based work at peak times falls further to
50–60% of employees, and office-based work at
non-peak times also falls to about 10%. Virtually all
this work transfers to home-based work at peak
times, accounting for about 20% of employees;
home-based work at non-peak times remains low.
These simulations suggest that environmental change
has much more impact on the activities of flexible
employees than those of unreformed employees.
Regarding the demand at the employer’s office shown
in Figures 3 and 4, the occupancy rate starts at effec-
tively 100% for unreformed employees in a traditional
environment; with flexible employees it drops to below
70% in traditional and intermediate environments, and
to below 60% in the modern environment – this last
scenario represents knowledge-based organizations.
In general, moving from the unreformed/traditional
scenario leads to graphs of occupation with a more flat-
tened shape, indicating longer periods of occupation
and lower numbers of occupants. However, even flex-
ible/modern scenario more work still takes place in
the employer’s premises than in employees’ homes.
The home-based working pattern of flexible employees
is highlighted in Figure 5. In the traditional
environment, it is concentrated in the time periods
before and after the usual working day; in the
modern environment the amount of home-based
work increases and the timing is inverted with most
taking place during the morning and afternoon time
periods. This phenomenon was not anticipated when
the model was developed and data values estimated;
it is an emergent property generated by the model.
Findings of the simulation model
Importance of both environmental and behavioural
change
The experiments explored the impact of changes to the
environment (performance and convenience scores)
and employees’ behaviour (work– life index). Chan-
ging both factors (moving from scenario U-T to F-M)
led to a significant impact on working patterns.
However, if only one factor was changed (from scen-
ario U-T to either F-T or to U-M) the impact was
much reduced. This implies that environmental and
behavioural changes are complementary aspects of
changing workstyles in the knowledge economy, and
that one or other in isolation has limited impact.
Demand at the employer’s premises
New workstyles reduce demand at the employer’s pre-
mises. The number of employees selecting this location
for work in the mornings and afternoons was lower
and also less stable. This creates an opportunity for
reducing the capacity of the employer’s premises.
However, with variable demand, if capacity is set at
the top end of the range, it will still be over-sized
most of the time. This might result in a move from
traditional premises with high capacity and high
utilization to modern premises with lower capacity
and lower utilization.
If capacity is reduced it is no longer feasible to assign
‘private’ workstations to individuals, as was possible
when most employees worked routinely at the employ-
er’s premises. The change to shared workstations
presents a challenge to premises managers.
Timing of home-based work
The remarkable transformation in flexible employees’
home-based working patterns, and non-peak hour
work at the employer’s office, suggest that evening,
night or dawn work is not desired but a response to
environmental constraints. In the modern environment
with fewer constraints (that is, higher performance
and convenience scores at both the home and office)
flexible employees are able to concentrate their work
activities in the morning and afternoon periods that
are conventionally preferred for work activities. This
implies that new workstyles unlikely to lead to round-
the-clock working. Indeed, an increase in out-of-hours
work by flexible employees in the traditional environ-
ment (F-T) compared with unreformed employees
Ta b l e 5 Six sc enarios used for the simulation modelling
experiments
Behaviour:
Environment
Unreformed Flexible
Traditional 1.Unreformed ^
traditional
4.Flexible ^
traditional
Intermediate 2. Unreformed ^
intermediate
5.Flexible ^
intermediate
Modern 3.Unreformed ^
modern
6.Flexible ^
modern
Modelling the use of space and time in the knowledge economy
319
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
(U-T) may not indicate is rising trend, but instead a tem-
porary phenomenon that disappears when the environ-
ment is modernized (F-M).
Comparison with survey data
The simulation model was based on a theoretical prop-
osition about activities in the knowledge economy.
Empirical confirmation would increase confidence in
the model and the findings.
The model findings can be compared with a survey of
modern organizations that was conducted as part of a
current research project titled ‘Environmental Impact
of Flexible Working’. In this project an on-line question-
naire was sent to nine commercial organizations in the
Figure3 Unreformed employees: work locations and times, and demand at the employers o⁄ce, for traditional, intermediate,and modern
environments: U-T, unreformed employees^traditional environment; U-I, unreformed employees^intermediate environment; U-M,
unreformed employees ^modern environment; HW, home work; and OW, o⁄ce work
Fawcett and Song
320
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
UK and two academic institutions,one in the UK and one
in the US, in May and August 2008. The analysis
reported here is based on 161 responses, 105 from com-
mercial companies and 56 from academia.Demographic
information is shown in Table 6.
Two survey questions are of interest. Question 5 asked
participants about the typical workplaces they used
during the week: ‘Please select times and places when
and where you work during a typical week’. Partici-
pants answered by selecting from drop-down menus
set out on a diary grid for the five sessions per day on
the five weekdays. The drop-down menus offered
seven alternatives: non-work, office, home working,
cafe
´, train, library, and other.
Question 6 asked about participants’ preferred work-
places: ‘Please select times and places when and
where you would PREFER to work. (Imagine you
can choose any time and place to work!)’. Participants
answered by selecting from the same drop-down
menus on a second diary grid.
The answers show that currently there is more home-
based work by participants in the evening and night
rather than the day. This pattern is similar to model
Figure 4 Flexible employees: work locations and times, and demand at the employers o⁄ce, for traditional, intermediate, and modern
environments: F-T, £exible employees^traditional environment; F-I, £exible employees^intermediate environment; F-M, £exible
employees^mo dern environment; HW, home work ; and OW, o⁄ce work
Modelling the use of space and time in the knowledge economy
321
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
scenario F-T (flexible agent– traditional environment).
However, the participants’ preferred workstyles
showed a move to home-based work during the day,
with a reduction in office-based work at this time.
This is similar to model scenario F-M (flexible
agent–modern environment). The survey results are
shown in Figure 6, averaged over all days in the
week. The typical and preferred levels of home-based
work over the whole week are shown in Figure 7.
Although the survey findings are limited in scope, they
are consistent with the results of the simulation model.
They also suggest that the organizations surveyed had
not yet fully adopted the potential of modern, knowl-
edge economy workstyles.
Future work
Model development
The preliminary simulation model offers many oppor-
tunities for enhancement.
Sensitivity analysis
The input data values were estimated and it would be a
useful exercise to undertake further model runs with a
wider range of input data values, possibly including
values that do not appear to be realistic. This would
improve understanding of how the output changes
with varying input data, allowing the model’s findings
to be interpreted with greater confidence.
Individuals or t ypes of individuals
In the simulations each individual employee is mod-
elled separately, but they are of just two types (unre-
formed and flexible). It would be possible for each
employee to have unique descriptive data. An inter-
mediate step would be to introduce a larger number
of employee types.
Unique descriptions would involve assigning each
employee a set of performance and convenience
scores, reflecting, for example, personal workstyles
and facilities for home-based work; and a set of
unique work–life index values, as opposed to relying
on random generation with the
s
(sigma) parameter.
Routine activity patterns
Work–life index values for each time period are gener-
ated without reference to the preceding or following
values, whereas in reality people organize their daily
activities in a sequence, and often form routines that
repeat from day to day. The model could be developed
to generate sequences and repeating cycles in individ-
uals’ work–life index values. However, the overall
pattern of variability between individuals may be
more significant than accuracy with which individuals’
diaries are modelled, if the model is used to assess
overall patterns of activity and demand.
Modes of flexible work
The model concentrates on home-based work, but
there are also other forms of flexible working, includ-
ing part-time work, flexitime, career breaks, annual-
ized hours, sabbaticals and unpaid leave. The model
of a week’s activities could be developed to allow for
part-time work and flexitime, but the others operate
on a longer timescale and would require a different
modelling approach.
Capacity constraint at an employer’s premises
There is no capacity constraint at the employer’s pre-
mises. In practice, if the demand approaches capacity,
employees would be expected to respond by changing
to home-based work or non-work. The model could
be developed to include a congestion-averse feedback
mechanism.
Travel-to-work time
No account is taken of employees’ travel to work
times, which are likely to affect decisions about
office- or home-based work. Changing the model to
take account of travel times would require significant
development.
Figure 5 Home-based working of £exible employees in
traditional, intermediate, and modern environments: F-T, £exible
employees^traditional environment; F-I, £exible employees^
intermediate environment; and F-M, £exible employees^modern
environment
Ta b l e 6 Demographic information of the sample
Sex Male¼49%, female¼51%
Age group B elow 20¼0.5%, 2 0s ¼29.4%, 30s¼30.2%,
40s¼15.5 %, 50 s¼11.7%, over 60¼12.7 %
Job
characteristics
Administration¼21.0%, marketing¼2.0%,
¢nance¼4.1%, engineer¼4.1% ,
operations¼4.4%, research¼32.4%,
teaching¼6.2%, design¼15.7%,
others¼10.1%
Fawcett and Song
322
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
Empirical comparison
Up to now the simulation model has relied on
estimated data values derived from published sources
and discussions with experts. An important step in
validation would be to test whether, by adjusting
model parameters (performance/convenience scores
and
s
(sigma)), the model can produce output corre-
sponding to survey observations in real organizations.
If this can be done successfully for several organiz-
ations, with the model parameters varying in a sys-
tematic way to reflect the different types of
organization, confidence in the model, and its fore-
casts, would be greatly increased.
Conclusions
These preliminary studies of new workstyles using an
agent-based simulation model have generated interest-
ing results. The significant indications are as follows:
.new workstyles depend on change to both beha-
viour and environment – if only one aspect is
changed the impact on working practices is limited
.new workstyles reduce demand at the employer’s
premises
.high levels of employee choice may reinforce the
established conventions about preferred working
times
Figure6 Work locations and times (data were collected in May and August 2008)
Figure7 Proportion of participants reporting home-based work
Modelling the use of space and time in the knowledge economy
323
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
The first indication is of greatest significance for man-
agers in organizations that are contemplating or under-
going change, emphasizing that environment and
behaviour are interdependent factors in the process of
change.
The second indication is familiar, but rarely quantified.
Estimates of the scale of demand are crucial infor-
mation for the briefing, design and management of
buildings for organizations in the knowledge economy.
The third indication questions the common assump-
tion that future activities will increasingly extend
over the 24-hour cycle: if validated, the model finding
would have direct implications for building manage-
ment, because the environmental impact of buildings
is much greater when the hours of operation are
extended, and it would be unnecessary to extend oper-
ating house in the absence of activity demand.
Although this is a preliminary study, it shows the value
of agent-based simulation for the investigation of
emerging or hypothetical scenarios – something that
cannot be done by observing current situations.
However, simulation model must be developed and
calibrated in parallel with empirical studies of current
situations, to give credibility to model findings for
hypothetical scenarios.
Simulation models are always simplified compared
with reality, and it is vital that the model captures
the key aspects of the system being studied, eliminating
only secondary or peripheral factors. The choice of
model structure is an implicit proposal regarding the
aspects that are believed to be significant. The model
described here is put forward as a contribution to an
on-going programme of research into the architectural
implications of the move to the knowledge economy.
References
Bonney, N. (2005) Overworked Britons? Part-time work and
work– life balance. Work, Employment and Society,19(2),
391–401.
Cousins, K.C. and Robey, D. (2005) Human agency in a wire-
less world: patterns of technology use in nomadic comput-
ing environments. Information and Organization,15,
151–180.
DEGW (2008) Working Beyond Walls: The Government Work-
place as an Agent of Change, Office of Government Com-
merce, London.
Department for Transport (DfT) (2005) Focus on Personal
Travel: 2005 Edition (available at: http://www.dft.gov.uk/
pgr/statistics/datatablespublications/personal/focuspt/
2005/).
Department of Business Enterprise and Regulatory Reform
(DBERR) (2007) Flexible Working and Work–Life Balance
(available at: http://www.dti.gov.uk/employment/work
andfamilies/flexible-working/index.html).
Duffy, F. (1992) The Changing Workplace, Phaidon, London.
Felstead, A., Jewson, N. and Walters, S. (2005) Changing Places
of Work, Palgrave, London.
Fitzner, G. and Grainger, H. (2007) The Right to Request Flexible
Working: A Review of the Evidence. Employment Relations
Research Series No. 59, Department of Trade and Industry,
London.
Halford, S. (2005) Hybrid workspace: re-spatialisations of work,
organisation and management. New Technology, Work and
Employment,20(1), 19–33.
Harrison, A., Wheeler, P. and Whitehead, C. (2003) The Distrib-
uted Workplace, Spon, London.
Hooker, H., Neathey, F., Casebourne, J. and Munro, M. (2007)
The Third Work–Life Balance Employee Survey: Main
Findings. Employment Relations Research Series No. 58,
Department of Trade and Industry, London.
Hyman, J., Scholaris, D. and Baldry, C. (2005) Getting on or
getting by: employee flexibility and strategies for home
and work. Work, Employment and Society,19(4),
705–725.
Lewis, S. and Cooper, C. (2005) Work–Life Integration: Case
Studies of Organisational Change, Wiley, Chichester.
Michelson, W. (2005) Time Use: Expanding the Explanatory
Power of the Social Science, Paradigm, Boulder, CO.
Michelson, W. and Crouse, D. (2004) Examining large-scale
time-use files through graphic representation. Electronic
International Journal of Time Use Research,1(1), 85–100,
(available at: http://www.eijtur.org/).
Myerson, J. and Ross, P. (2006) Space to Work: New Office
Design, Laurence King, London.
Robert, M. and Borjesson, R. (2006) Company incentives and
tools for promoting telecommuting. Environment and Beha-
vior,38(4), 521–549.
Strelitz, Z. and Edwards, M. (2006) Getting It Together: The
Work–Life Agenda and Offices, British Council for
Offices, London.
Worthington, J. (2005) Reinventing the Workplace, 2nd edn,
Architectural Press, Oxford.
Fawcett and Song
324
Downloaded By: [Lorch, Richard] At: 10:28 7 May 2009
