In travel behaviour modelling, latent class models are used to represent underlying discrete groupings of behavioural preferences. The paper presents a latent class extension of a dynamic discrete choice model (DDCM) and applies the model to the problem of activity demand generation and scheduling. The DDCM is a recursive multinomial logit model where agents make sequential decisions in time, maximizing the expected future utility of their decisions in a random utility maximization framework. It generates activities and their associated travel within a full day schedule, endogenously respecting agents' inherent time-space constraints. The latent class DDCM builds on the base model by representing heterogeneous lifestyle preferences. A specification of the model is estimated on a Stockholm travel survey and uses age, income level, gender, car ownership and presence of children in the household as classifying variables. The models result in classes which primarily represent modality styles, finding car-, transit- and bike-primary behavioural groups as well as a multimodal group, each linked with different socio-demographic characteristics. The models improve over non-latent class reference models and provide insight into the structure of heterogeneity in travel behaviour preferences in Stockholm.

Crowdshipping has grown in popularity as a sharing economy model, but ensuring its sustainability remains a challenge. This study explores how public transit riders can be engaged in crowdshipping services to avoid generating additional motorized traffic. The propensity of public transit users to participate in crowdshipping and their responses to alternative task attributes are explored through an in-person survey conducted at selected subway stations in Stockholm. The influence of different socio-demographic factors and trip features on the propensity for participation is examined using statistical analysis and regression models. To quantify the trade-offs among required detours, compensation, and parcel weight when accepting crowdshipping tasks, alternative discrete choice models are investigated. The results reveal that factors such as age, employment, and income, along with trip characteristics, significantly affect participation propensity. The estimated willingness to work as a crowdshipper aligns with previous studies showing that age and income level were important factors. A latent class model further reveals a clear division between two groups: one younger, lower income group with higher willingness to work, and another older, higher-income group with lower willingness. As a result, dedicated strategies need to be considered by future crowdshipping service providers and policymakers.

The study adapted the Perceived Accessibility Scale to assess perceived cycling accessibility in Stockholm, presenting a new case study of its application. Using recent survey data and factor analysis, the scale demonstrated strong reliability in capturing perceptions. Hypothesis testing indicated limited effects of most sociodemographic factors on perceived cycling accessibility, except for gender among frequent cyclists. Mobility-related characteristics and peer influence had stronger effects, while spatial variables were insignificant. A significant positive association was also found between perceived cycling accessibility and cycling frequency. These findings align with expectations and reflect behaviour of Swedish cyclists.

Energy efficient control of energy systems in buildings is a widely recognized challenge due to the use of low temperature heating, renewable electricity sources, and the incorporation of thermal storage. Reinforcement Learning (RL) has been shown to be effective at minimizing the energy usage in buildings with maintained thermal comfort despite the high system complexity. However, RL has certain disadvantages that make it challenging to apply in engineering practices. In this review, we take a computer science approach to identifying three main categories of challenges of using RL for control of Building Energy Systems (BES). The three categories are the following: RL in single buildings, RL in building clusters, and multi-agent aspects. For each topic, we analyse the main challenges, and the state-of-the-art approaches to alleviate them. We also identify several future research directions on subjects such as sample efficiency, transfer learning, and the theoretical properties of RL in building energy systems. In conclusion, our review shows that the work on RL for BES control is still in its initial stages. Although significant progress has been made, more research is needed to realize the goal of RL-based control of BES at scale.

Den här rapporten utgör slutrapport för projektet Bilrestider i storstad: Variationsmönster och upplevd osäkerhet (VARIA) som genomförts på uppdrag av Trafikverket (TRV 2018/16380).Det övergripande syftet med det här projektet har varit att vidareutveckla metoder för samhällsekonomiska kalkyler för olika åtgärder i trafiksystemet. Projektet bidrar till detta mål genom att utveckla kunskap om hur olika komponenter i restidsvariation påverkar trafikanternas faktiska erfarenhet av restidsvariation, och trafikanternas upplevelse av systematisk restidsvariation och deras restidsosäkerhet. 

Forskningsfrågorna studeras i två delstudier. En delstudie fokuserar på trafikanternas förståelse av restidsvariation, och deras upplevda restidsosäkerhet. Denna delstudie baseras på ett teoretiskt ramverk och två datakällor: (1) en enkätstudie av hur tillfrågade bilister beskriver den restidsfördelning de förväntar sig (och därmed kan antas planera för), samt (2) empiriska data som beskriver hur den verkliga restiden längs samma rutter varierar i olika dimensioner.

Den andra delstudien fokuserar på sambandet mellan hur restiden varierar i data som uppmätts generellt, och den restidsvariation som enskilda bilister faktiskt utsätts för. Här utnyttjas ett statistiskt ramverk och en tredje datakälla: restidsobservationer från olika rutter på enskild passagenivå med beständiga fordons-ID.

En grundläggande svaghet med den första delstudien ligger i att respondenterna i sina uppgivna restider inte tycks ha avgränsat sig på det sätt som var avsikten (bara rena körtider mellan de uppgivna ändnoderna). Trots detta anser vi att det utvecklade analytiska ramverket gör det möjligt att dra vissa övergripande slutsatser med bäring på studiens inledande frågeställningar.

Resultaten från den första delstudien stöder sammantaget hypotesen att resenärerna i gemen i huvudsak baserar sin planering och schemaläggning inför en specifik resa på mer generella restidsprediktioner, som är underbyggda av deras samlade erfarenhet, (snarare än specifika erfarenheter eller annan information om restiden på den specifika rutten). Resultaten visar också att förare som kan bygga under sinna uppskattningar med mer specifik erfarenhet, har minst lika svårt att förutse hur restiden varierar, såväl när det gäller ”systematisk” som ”slumpmässig” variation, som förare som saknar specifik erfarenhet från just den rutt för vilken restiden skall uppskattas. 

Respondenternas prediktioner av restidsvariation är sämre än deras prediktioner av förväntad restid. Detta stöder hypotesen att restidens faktiska variation för en viss specifik rutt, under en viss specifik tid på dagen ger en långt ifrån komplett bild av den restidsosäkerhet som resenärer måste ta höjd för i sin planering och schemaläggning. Om samhällsekonomiska värderingar av restidsosäkerhet tillämpas direkt på uppmätt (eller predicerad) restidsvariation, utan hänsyn till det komplexa sambandet dem emellan, kan värdet av minskad restidsvariation såväl över- som underskattas. 

I den andra delstudien har vi undersökt i vilken utsträckning fordonen som korsar en rutt är återkommande resenärer, och hur den andelen beror av olika attribut. Med hjälp av data från Bluetooth- och Wifi-sensorer under en tremånadersperiod har vi funnit att det genomsnittliga antalet resor per fordons-ID är högre mot staden på morgontoppen och ut från staden på eftermiddagen, vilket är förenligt med vetskapen att pendlingsresor tenderar att ha den högsta regelbundenheten över dagar. Vi har även föreslagit en modell för hastighetsfördelningar hos rutter, som separerar variationen i en komponent med variation mellan resenärer, och en komponent med variation inom varje resenär (”individuell erfarenhet av restidsvariation”). Resultaten av modellberäkningar visar att den relativa individuella (inom resenären) variationen är betydligt högre i pendlingsriktningen (mot staden på morgonen och ut från staden på eftermiddagen) och på rutter med hög trängselnivå. Trängsel tycks alltså vara den viktigaste faktorn som förklarar den relativa restidsosäkerheten. 

På grund av en viss omnumrering över tiden av fordons-ID i datan som använts är den exakta frekvensen med vilken resenärer använder en rutt samt storleken på variabiliteten mellan resenärer och inom varje resenär svåra att skatta. Den beräknade frekvensen är låg, vilket i och för sig är i linje med analyser av data från trängselavgiftsportaler. Eftersom omhashningen rimligen är oberoende av geografi och tid på dagen så har vi kunnat göra relativa studier mellan rutter och tidsperioder. Resultaten indikerar att det måste göras en åtskillnad mellan den variabilitet som en enskild resenär kan ha erfarenhet av, och den totala variabiliteten som vanligtvis används i bedömningarna av restiders tillförlitlighet. Den relativa storleken på de två termerna varierar systematiskt med ruttegenskaper och tidsperioder. Utan denna åtskillnad kan kostnaderna för restidsvariabilitet överskattas

This paper is the second part of a two-part series that investigates the energy-saving potentials of a novel wastewater heat recovery system connected to mechanical ventilation with heat recovery (MVHR). The main idea is to use the heat from stored wastewater to preheat the incoming cold outdoor airflow to the MVHR and thereby reduce the defrosting needs of the MVHR. The study evaluated the potential of two air preheaters, AP 0.4 m x 0.4 m x 0.4 m and AP 0.8 m x 0.8 m x 0.4 m , placed in front of the existing MVHR. Dynamic simulations in this study have shown that the smaller air preheater could lower the frost threshold temperature by 5 °C and the larger one could reduce it by 11 °C. Without an air preheater, the defrosting was needed during nearly two-thirds of January in the studied climate. By contrast, with the evaluated air preheaters in front of the MVHR, the defrosting was needed during 45% and 20% of the evaluation period, respectively. The results also showed that frost growth inside the heat exchanger could be reduced by 38% with AP 0.4 m x 0.4 m x 0.4 m and by 62% with AP 0.8 m x 0.8 m x 0.4 m during the peak load. The main conclusion is that the suggested heat recovery system has good potential for improving the overall performance of MVHR systems in cold climates.

Energy and environmental issues are increasingly important in existing building service and energy systems around the world. Despite great efforts to implement retrofit techniques in Sweden, no stringent evaluation of the benefits of these techniques or their systematic design has been completed. Traditional evaluations have not taken into account the embodied energy and greenhouse gases emissions of different retrofit options. This omission leads to underestimation of the potential environmental benefits of modern retrofit techniques. In this study a novel hybrid modeling approach to quantify the sustainability of retrofit options is developed to fill these knowledge gaps. The compatibility of environmental and energy saving modeling of various energy-saving techniques for future transition of Swedish residential building stock is analyzed. Consolis Retro and the life cycle assessment (LCA) techniques are employed and further coupled to simulate retrofit options. The model integrates both energy demand (net operational energy), primary energy (operational energy from energy mix to buildings) into evaluation criteria. Embodied energy (energy required to produce materials of retrofitting options) and embodied greenhouse gas emissions (upstream CO<inf>2</inf> equivalent) are introduced as new measures in the evaluation criteria. The results showed that low-temperature heating retrofitting was the most effective option from both a primary and embodied energy perspective in the studied building types. Combining circulation pump renovations could further contribute to the efficiency of low-temperature heating for energy-demand savings. High operational energy-saving measures may not always lead to larger reduction in both embodied energy and greenhouse gas emissions, particularly for building envelope retrofitting. Neglecting the embodied energy of retrofit options will increase the risk of overrepresenting their energy-saving contributions. The sustainability improvements of retrofitting, particularly large-scale measures, should take into account the embodied energy and greenhouse gas emissions from the material productions.

Urban geography could be characterized by analysing the patterns that describe the flows of people and goods. Measuring urban structures is essential for supporting an evidence-based spatial planning policy. The objective of this study is to examine how the spatial-temporal distribution of public transport passenger flow could be used to reveal urban structure dynamics. A methodology to identify and classify centres based on mobility data was applied to Metropolitan Stockholm in Sweden using multi-modal public transport passenger flows. Stockholm is known for its long-term monocentric planning with a dominant central core and radial public transport system. Strategic nodes along its radial public transport system have been a focus for development of sub-centres. Although the regional planning policy embraces a shift towards a polycentric planning policy, the results indicate that this has not been realized insofar.

Scheduling preferences, derived under the assumption of expected utility maximization, can be used for measuring the value of travel time variability. As an alternative to the expected utility maximization, the rank dependent utility is chosen in this paper. Considering a flexible probability weighting for ranked outcomes, two frequently applied scheduling models are derived and estimated on a data set with specific travel time distributions. Optimistic respondents are observed. Allowing for heterogeneity of scheduling preferences improves the model performance significantly. We also found evidence for valuing excessive travel time, which is not included in the traditional scheduling model specifications.