CHEMICAL ENGINEERING TRANSACTIONS

VOL. 81, 2020

A publication of

The Italian Association
of Chemical Engineering
Online at www.cetjournal.it

Guest Editors: Petar S. Varbanov, Qiuwang Wang, Min Zeng, Panos Seferlis, Ting Ma, Jiří J. Klemeš

ISBN 978-88-95608-79-2; ISSN 2283-9216

Optimal Location for Conference Venue in Relation to

Transport Emission Sustainability Strategy

Vlastimír Nevrlýa,*, Yee Van Fanb, Radovan Šomplákb, Veronika Smejkalováa, Jiří

Jaromír Klemešb, Hon Loong Lamc

aInstitute of Process Engineering, Faculty of Mechanical Engineering, Brno University of Technology, Technická 2896/2,

616 69 Brno, Czech Republic
bSustainable Process Integration Laboratory – SPIL, NETME Centre, Faculty of Mechanical Engineering, Brno University of

Technology – VUT Brno, Technická 2896/2, 616 69 Brno, Czech Republic
cCentre of Excellence for Green Technologies, University of Nottingham Malaysia Campus, Jalan Broga,43500 Semenyih,

Selangor, Malaysia

Vlastimir.Nevrly@vutbr.cz

Thousands of conferences held all over the world annually are closely related to the theme of environmental

sustainability. Various methods and measures have been presented in the effort to reduce environmental

impacts. However, the impact of the conference itself receives relatively less attention. This study proposed a

mathematical model to identify the optimal conference locations with minimal environmental footprints in

travelling to and from the conference by the participants. The historical data of Conference on Process

Integration for Energy Saving and Pollution Reduction (PRES) consisting the origin of the participants is applied

to demonstrate the location selection. The model is developed to include air transport for short and long

distances where the environmental impact is considered in the form of external cost. Participants are

geographically distributed into points considering the capital of countries. Socio-economic and geographic data

are used to estimate the expected number of participants through a linear regression model. Population, the

absolute difference in religion, the distance between countries (with a positive impact on total attendance) and

continent conservation (negative impact) were found to be significant. The result suggests the Czech Republic

as the optimal location with the external cost of about 242 € / cap. The subsequent locations are Estonia and

Romania. Future studies can cover multi-objective decision making, including price and more problematically

quantified features as safety, health issues and attractiveness.

1. Introduction

Conferences bring researchers from different geographical areas together at a place for learning the latest

research, exchanging the knowledge and provide ample opportunities for networking. Environmental

sustainability is one of the most discussed themes or topics in a conference; however, the own environmental

footprint of the conference is yet to be responded (Caset et al., 2018). Holden et al. (2017) found that most of

the conference organiser did not think about the environmental impacts of their meeting. Jäckle et al. (2019)

estimated that the average emissions per attendee of European Consortium for Political Research general

conference were between 0.5 - 1.3 t CO2eq / cap. The greenhouse gases (GHG) footprint is significant

considering the high share by taking the value of IPCC where a reduction to 2.5 t CO2eq / cap by 2030 and 0.7 t

CO2eq / cap by 2050 is required to keep on track with 1.5 oC goal. Neugebauer et al. (2020) addressed scientific

conferences as a source of environmental burdens, with 0.57 t CO2eq / cap. However, it should also be noted

that, without the conferences, the environmental footprint would not be necessarily considerably lower. More

studies are still needed for verification if the business travel is replaced by leisure travel or other activities which

are still inducing GHG footprint. The transportation and hotel are still running sometimes with even a higher

CO2eq / cap (e.g. when the capacity is not fully utilised). The differences in CO2eq / cap compared to the baseline

scenario would provide a better insight rather than referring to the absolute value. There has been an increasing

call for sustainable conferencing (Bossdorf et al., 2010). Some of the efforts included the implementation of the

DOI: 10.3303/CET2081117

Paper Received: 01/04/2020; Revised: 28/04/2020; Accepted: 02/05/2020
Please cite this article as: Nevrlý V., Fan Y.V., Šomplák R., Smejkalová V., Klemeš J.J., Lam H.L., 2020, Optimal Location for Conference
Venue in Relation to Transport Emission Sustainability Strategy, Chemical Engineering Transactions, 81, 697-702 DOI:10.3303/CET2081117

697

paperless conference (distributing electronic materials instead of the printed book) and replacement of plastic

cups with paper cups, glasses or bottles. Teleconference/virtual conference has also been proposed to minimise

the environmental footprint. However, in enhancing the environmental performance, the effectiveness of the

meetings through in-person interaction and discussion (networking) should not be compromised entirely.

According to the regression analysis by Mair and Thompson (2009), face to face networking opportunities are

one of the significant underlying dimensions in attending a conference.

Desiere (2016) stated that the footprints of non-European participants in a European conference are significantly

high, accounted for > 50 % of the total emissions, with only 10 % of total participants. This reflects the importance

of selecting the conference venue according to the distribution of participants. Identification of an optimised

location for conferences can mitigate the environmental footprints, but it is not a straightforward optimisation. It

involves the trade-offs between different factors on top of the environmental concern. Jo et al. (2019) assessed

the location selection process and attributes for international association meetings where more than 24 crucial

attributes are identified. Available studies on location optimisation usually focus on industry processes and

waste management. It has not been properly applied to the study of the conference venue.

This study intends to identify an optimal location for conference organising focuses on environmental aspects,

by considering the nature (e.g. distribution of participants) of the conference.

The novel contributions of the presented study are

• The consideration of the different type of emissions in defining environmental performance and selecting

an optimal location.

• An improved mathematical model to identify the conference venue based on the regression model from

the historical records.

• A demonstration by a real case study where the results could provide higher practical value – suggestions

for future conference venues with least environmental footprint.

2. Problem description and simplifications

The core of the proposed problem is the environmental impact of transportation to and from conferences of all

participants. It strongly depends on the origin of participants and conference venue – selected location. The

total participant’s environmental impact should be calculated throughout the whole journey from his home,

through some inter-connections, to the final destination. In particular, it may be composed of multiple and

different modes of transport (tram, bus, train, boat, aircraft). An exemplary route may be given, starting at home,

taking the short tram and then underground to the central train station, where a shuttle bus to an international

airport is used. After the flight, shuttle buses/taxis are used to get to the city centre, where local public transport

is used to get to the hotel (close or directly at the conference venue). This example might be further extended

with inter-connection flights and transfers. Each of the means of transport used relates to some energy, which

has to be generated. Depending on the mean of transport, the respective emissions are produced and emitted

to the environment. Even tram or underground which are powered by electricity have some impact (depends on

the source of energy and respective emission production) unless it is based on cleaner energy (wind, water).

The problem is formulated as where to hold a conference to have a minimal environmental impact caused by

necessary participants’ transportation to and from the venue. In the first step, the problem is simplified, and only

air transportation is considered. It is likely that different results can be obtained when the train and other transport

modes are taken into account. Each of the historical participants is represented by an international airport

situated at the capital city of his state of origin. Direct flights are then assumed to all candidate conference venue

locations. A notable simplification is underlying in this assumption, because large countries may benefit from

small ones. For example, China, with almost 10,000,000 km2, and a huge number of large cities, is considered

in a single point as also the Czech Republic with less than 80,000 km2. This means that all participants from

one country are represented by only one point. Researchers attending the conference should begin their journey

from their university, with the possibility to choose from various available international airports. However, the

problem presented here reflects only one centrally selected international airport.

The considered airports corresponding with five past conference participants’ country of origin are depicted in

Figure 1. Red labelled marks denote the international airport assigned to recent venues. This network of airports

and respective distances between them can be used for the calculation of optimal conference location. Historical

participation data is required to calculate a suitable location. However, these are only available for venues where

the conference has already taken place. In some places, it is a traditional meeting of the academic community

and is repeated regularly. Since for many candidate locations, no data of attendance are available; it is

necessary to make an estimate for the potential number of participants in each country considered. The origin

of participants is key for the calculation of environmental impact. This work focus on the first step of the elaborate

assessment approach.

698

Figure 1: Illustration of the problem and the scope of the proposed approach. Red labels indicate the Czech

Republic, Italy, Greece, China and Malaysia, where the previous conferences were located.

3. Mathematical model

Emissions from air transport are a major contributor to GHG production (Amizadeh et al., 2016). One of the

parameters that affect specific values is the type of aircraft, but also the age of aircraft and time in service

influence fuel consumption and emission. The use of the type of aircraft depends on the distance of the flight

and the territorial classification of the transport (internal / external traffic). Depending on travelled distance, the

aircraft produces a different amount of emissions. Two options were decided to be included in the model

corresponding with the short and long haul. Integer linear programming techniques are used to formulate the

model. The notation of the model attributes is described in Table 1.

Table 1: The notation

Type Symbol Description

Sets 𝑖 ∈ 𝐼 Nodes of the network representing participants University origin

𝑗 ∈ 𝐽 Nodes defining candidate locations for the conference venue

𝑝 ∈ 𝑃 Set of pollutants

Parameters 𝑛𝑖,𝑗 Number of participants from node 𝑖 attending location 𝑗

𝑑𝑖,𝑗 Distance from participant’s node 𝑖 to candidate location 𝑗

𝑒𝑖,𝑗,𝑝 Pollutant 𝑝 produced by transport from 𝑖 to 𝑗

𝑎𝑝
𝑆, 𝑎𝑝

𝐿 Pollutant production for short and long distances

𝑡 Threshold dividing short and long-distance

𝑐𝑝 Metric for pollutant’s 𝑝 impact

𝑀 Upper bound – total number of candidate sites 𝑗

Variables 𝛿𝑖,𝑗 Binary variable

It is assumed that emitted pollutants depend on the travelled distance. The shorter the trip is, the higher is the

production of emission per kilometre. The pollutants’ production is considered as in Eq(1), where the threshold

distance 𝑡 plays the distribution point.

𝑒𝑖,𝑗,𝑝 = {𝑎𝑝
𝑆𝑑𝑖,𝑗 𝑖𝑓 𝑑𝑖,𝑗 ≤ 𝑡, 𝑎𝑝

𝐿 𝑑𝑖,𝑗 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒, ∀𝑖 ∈ 𝐼, ∀𝑗 ∈ 𝐽, ∀𝑝 ∈ 𝑃. (1)

The model of the described problem is formulated by Eq(2)-Eq(6). The objective function Eq(2) summarises the

impact of all considered pollutants produced by the transportation of participants to the conference location.

Multiplier 2 stands for the return journey. The equation is normalised by the number of participants attending

the respective location so that it is comparable. The crucial part of the objective function is the parameter 𝑐𝑝,

which defines the selected criterion for pollutant impact.

∑

𝑖∈𝐼

∑

𝑗∈𝐽

∑

𝑝∈𝑃

𝑐𝑝𝑒𝑖,𝑗,𝑝𝑛𝑖,𝑗 2𝛿𝑖,𝑗
∑ 𝑘∈𝐼 𝑛𝑘,𝑗

(2)

The constraints take the following form:

∑

𝑗∈𝐽

𝛿𝑖,𝑗 = 1 ∀𝑖 ∈ 𝐼, (3)

699

∑

𝑖∈𝐼

𝛿𝑖,𝑗 ≤ 𝑀𝛾𝑗 ∀𝑗 ∈ 𝐽, (4)

∑

𝑗∈𝐽

𝛾𝑗 = 1 (5)

𝛿𝑖,𝑗 ∈ {0,1}
∀𝑖 ∈ 𝐼, ∀𝑗 ∈ 𝐽.

(6)

Eq(3) ensures that for all candidates, a single location is selected. In Eq(4), the choice of venue locations is

activated, while Eq(5) specifies the number of conference venues. Eq(6) states variables as binary. The goal of

the presented model is to minimise the environmental impact per participant. The distance 𝑑𝑖,𝑗 between airports

is calculated based on the GPS coordinates and the following Eq(7), which defines the distances on the sphere.

Parameter 𝑟 is the sphere radius and [ 𝜓𝑖 ; 𝜆𝑖 ], [ 𝜓𝑗 ; 𝜆𝑗 ] are coordinates of the points (airports). The distance

calculation is simplified and could be further edited according to real air routes.

𝑑𝑖,𝑗 = 𝑟 ⋅ 𝑎𝑟𝑐𝑐𝑜𝑠 (𝑠𝑖𝑛 𝑠𝑖𝑛 𝜓𝑖 𝑠𝑖𝑛 𝑠𝑖𝑛 𝜓𝑗 +𝑐𝑜𝑠 𝑐𝑜𝑠 𝜓𝑖 𝑐𝑜𝑠 𝑐𝑜𝑠 𝜓𝑗 𝑐𝑜𝑠 𝑐𝑜𝑠 (𝜆𝑗 − 𝜆𝑖 ) ) (7)

The presented simple model demonstrates the choice possibility of conference venue location from an

environmental point of view, where only air transport is considered. Based on obtained results related to the

most important parameters, it could be enhanced in order to reflect reality more credibly. Extension options are

transfers possibility, new infrastructure (network) corresponding with road and/or rail transportation, and densely

designed network of nodes (participants’ university / city, country division, multiple airports).

4. Case study – PRES conference

The mathematical model is applied to a real case study to suggest a location for the upcoming PRES

conferences. The estimation is based on the participation data of PRES conference (Klemeš et al., 2017), which

is having a history for more than 20 y. A total of 72 countries which have attendance history were considered

as candidate locations for the venue. The transportation mode in this case study is limited to air transport (plane),

where the major airport of the countries’ capital is set as the reference point. The considered aviation emissions

include CO2, CH4, N2O, NOX and SOX, and the emission factors applied are in Table 2. Emission factors are

divided with a threshold distance of 463 km, according to LISPASTO (2017) in defining short distance and long-

haul flight. Values in Table 2 are normalised per passenger and km. The total external cost has been selected

to represent the environmental impacts and as the criterion in the objective function to be minimised. It is

calculated according to Eq(9), where 𝑐𝑝 is the cost coefficient of the emission type (e.g. CO2, CH4, N2O, SO2,

NOX) in €/t and value 𝑒𝑝 is the emissions amount (e.g. in g or t). Variables 𝑒𝑝 and 𝑐𝑝 correspond with 𝑒𝑖,𝑗,𝑝 and

𝑐𝑝 in Eq(2). Considered values of environmental prices are in Table 2.

Table 2: Emission factors depending on distance (LIPASTO, 2017) and environmental prices (CE Delft, 2017)

Emission Type Short Distance (<463 km) [ g / (p×km)] Long Haul [g / (p×km)] Environmental Price [ € / t]

CO2 257 113 56.6

CH4 0.0018 0.00078 175

N2O 0.0070 0.0031 15,000

NOX 1.1 0.45 34,700

SOX 0.084 0.036 24,900

𝑇𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙𝑐𝑜𝑠𝑡 = ∑

𝑝∈𝑃

𝑒𝑝𝑐𝑝 (9)

A linear regression model was used to estimate the number of participants for all candidate locations. The

chosen independent variables are gross domestic product, population, the distance between countries

(airports), percentage distribution of faith, the absolute difference in religion, the number of tourist per capita

and continent conservation. Historical data regarding the number of participants served as the dependent

variable. Participants from the organising country were excluded from the regression model. Data of factors

were taken from the database of the World Bank Open Data (World Bank Group, 2018). According to the

formulated model, only four independent variables were found significant, according to the p-value supporting

the null hypothesis. The resulting linear model includes population (𝑃𝑖), the absolute difference in religion (𝑅𝑖,𝑗),

the distance between countries (𝑑𝑖,𝑗 ) and continent conservation (𝐶𝑖,𝑗 ), while it describes 45 % of the variability

700

in data. The expected number of participants from country 𝑖 to candidate location 𝑗 is estimated, according to

Eq(8).

𝑛𝑖,𝑗 = 1.57788 ⋅ 10
−8𝑃𝑖 + 4.744224171𝑅𝑖,𝑗 + 0.00077018𝑑𝑖,𝑗 − 6.316886133𝐶𝑖,𝑗 (8)

The regression model indicates that the size of the population has a positive effect on the number of participants

which was expected. The difference in religion and distance increases the number of participants, which can be

interpreted as a location attractiveness for the participants. It was found that participants are less willing to travel

outside their continent due to the conference. The model serves for demonstrative purposes, for further and

more detailed analysis, it can be refined using new variables (attractivity, interactions, etc.) and adding data

from other conferences. The linear model Eq(8) was used to generate input data for the optimization problem.

5. Results and discussion

All stated models and equations were used to calculate the optimal location for the conference venue. The

optimal location based on external cost per participant is the Czech Republic (with capital city Prague) with an

objective function value of 242 € / cap. An illustrative transportation map is depicted in Figure 2. The map shows

the number of participants and total external cost from each continent. These values would be saved in the case

of remote presentation where no travelling is needed.

Figure 2: Results of the optimal conference venue location (EC=external cost, PC=number of participants)

The external costs were calculated for locations of past five PRES conferences, see Table 3 for comparison.

Table 3: Comparison of outputs for the past five PRES conference locations

Country Estimated number of participants External cost [ € / p] Total external cost [€]

China 738 354 260,857

Czech Republic 586 242 141,776

Greece 406 262 106,243

Malaysia 760 425 322,875

Italy 384 283 108,583

They were organised in Greece in Crete and Rhodes (2019, 2013), the Czech Republic in Prague (2018, 2016,

2014, 2012), China in Tianjin (2017) and Malaysia in Kuching (2015). The calculation is based on the assigned

international airport. The table includes countries, the expected number of participants and calculated external

costs (based on the estimated number of participants and emission produced). External costs were calculated

in total and also per participant for comparison with the result of the optimisation model (both values are

rounded) corresponding with the Czech Republic. If this location is banned, the subsequent optimum location

with minimal cost would be Estonia (Romania afterwards). These countries do not have many international long-

haul flights and most delegates would fly via Germany, UK, or Paris in reality. When total external costs are

assumed in the objective function instead of normalised by the number of participants, the suggested location

would be Ukraine with expected 380 participants and external cost equal to 96,686 €. Such a value is caused

by the low number of participants expected. The limitation of the regression model is that it underestimates the

number of participants from organising countries. A more accurate regression model would lead to different

results. The suggestion is to model local and foreign participants separately China could be suggested instead

as an optimal location thanks to the high number of locals. Specifying the different modes of transport and the

entire transport chain would also increase the credibility of the presented approach. Another model could be

701

developed for participants from the organising country to estimate the total attendance more accurately. Worth

discussion is also the choice of flight, which depends on the participant. This is often decided on the basis of

the price of air tickets and may lead to a transfer flight preference causing higher impact. The final decision

should also consider the maximisation of participants’ interactions which are also influenced by the diversity of

countries. Such a measure could be done by estimating the number of universities connected to the conference.

6. Conclusion

The presented approach suggested holding the PRES conference in the Czech Republic, which offers the

lowest environmental footprints in transporting. The estimated external cost is 242 € / cap. However, the result

is limited to air transportation and other assumptions as discussed. A sequential computation can propose an

order for the conference venues for several years. The subsequent conferences would be in Estonia and

Romania, with 243 € and 253 € / cap. The conference venue is also dependent on contacts of organisers with

the local academic community (these could account up to 50 % of attendance) where the list of candidate

locations can be restricted. The methods proposed can be used for a general event, while new factors would

be included. The model can also be formulated as multi-objective to take into account other essential criteria

that play a role in organising a conference. Essential factors to be considered for future study are social factors,

attractiveness, price and safety. The multi-criteria in decision-making may incorporate ranking approach. The

scope can also be extended by quantifying other environmental impacts such as the average carbon footprint

of PRES attendance.

Acknowledgements

The authors gratefully acknowledge the financial support provided by the project Sustainable Process

Integration Laboratory – SPIL, funded as project No. CZ.02.1.01/0.0/0.0/15_003/0000456, by Czech Republic

Operational Programme Research, Development, and Education, Priority 1: Strengthening capacity for quality

research and by ERDF within the research project No. CZ.02.1.01/0.0/0.0/16_026/0008413 "Strategic

Partnership for Environmental Technologies and Energy Production".

References

Amizadeh F., Alonso G., Benito A., Morales-Alonso G., 2016, Analysis of the recent evolution of commercial air

traffic CO2 emissions and fleet utilization in the six largest national markets of the European Union, Journal

of Air Transport Management, 55, 9-19.

Bossdorf O., Parepa M., Fischer, M., 2010, Climate-neutral ecology conferences: just do it!, Trends in Ecology

& Evolution, 25(2), 61.

Caset F., Boussauw K., Storme T., 2018, Meet & fly: Sustainable transport academics and the elephant in the

room, Journal of Transport Geography, 70, 64-67.

CE Delft, 2017, Environmental prices handbook <www.cedelft.eu/en/environmental-prices> accessed

26.03.2020.

Desiere S., 2016, The carbon footprint of academic conferences: Evidence from the 14th EAAE Congress in

Slovenia, EuroChoices, 15(2), 56-61.

Holden M. H., Butt N., Chauvenet A., Plein M., Stringer M., Chades I., 2017, Academic conferences urgently

need environmental policies, Nature Ecology & Evolution, 1(9), 1211-1212.

Jäckle S., 2019, WE have to change! The carbon footprint of ECPR general conferences and ways to reduce it,

European Political Science, 18(4), 630-650.

Jo D., Park H. Y., Choe Y., Kim D. K., 2019, Destination-selection attributes for international association

meetings: A mixed-methods study, Journal of Destination Marketing & Management, 13, 61-72.

Klemeš J.J., Varbanov P.S., Fan Y.V., Lam H.L., 2017, Twenty years of PRES: Past, present and future -

Process Integration towards sustainability, Chemical Engineering Transactions, 61, 1-24.

LISPASTO, 2017, Unit emissions <lipasto.vtt.fi/yksikkopaastot/indexe.htm> accessed 02.03.2020.

Mair J., Thompson K., 2009, The UK association conference attendance decision-making process, Tourism

Management, 30(3), 400-409.

Neugebauer S., Bolz M., Mankaa R., Traverso M., 2020, How sustainable are sustainability conferences?

Comprehensive Life Cycle Assessment of an international conference series in Europe, Journal of Cleaner

Production, 242, 118516.

World Bank Group, 2018, World Bank Open Data <data.worldbank.org/> assessed 13.02.2020.

702