(51f) A Personal Exposure Agent Based Model (ABM) with the Capacity of Aggregation at Various Levels of Population Size

/* Copyright 2012 Mozilla Foundation Copyright 2013 Lu Wang Apachine License Version 2.0 */ (function(){function b(a,b,e,f){var c=(a.className||"").split(/\s+/g);""===c[0]&&c.shift();var d=c.indexOf(b);0>d&&e&&c.push(b);0<=d&&f&&c.splice(d,1);a.className=c.join(" ");return 0<=d}if(!("classList"in document.createElement("div"))){var e={add:function(a){b(this.element,a,!0,!1)},contains:function(a){return b(this.element,a,!1,!1)},remove:function(a){b(this.element,a,!1,!0)},toggle:function(a){b(this.element,a,!0,!0)}};Object.defineProperty(HTMLElement.prototype,"classList",{get:function(){if(this._classList)return this._classList; var a=Object.create(e,{element:{value:this,writable:!1,enumerable:!0}});Object.defineProperty(this,"_classList",{value:a,writable:!1,enumerable:!1});return a},enumerable:!0})}})(); (function(){/* pdf2htmlEX.js: Core UI functions for pdf2htmlEX Copyright 2012,2013 Lu Wang and other contributors https://github.com/coolwanglu/pdf2htmlEX/blob/master/share/LICENSE */ var pdf2htmlEX=window.pdf2htmlEX=window.pdf2htmlEX||{},CSS_CLASS_NAMES={page_frame:"pf",page_content_box:"pc",page_data:"pi",background_image:"bi",link:"l",input_radio:"ir",__dummy__:"no comma"},DEFAULT_CONFIG={container_id:"page-container",sidebar_id:"sidebar",outline_id:"outline",loading_indicator_cls:"loading-indicator",preload_pages:3,render_timeout:100,scale_step:0.9,key_handler:!0,hashchange_handler:!0,view_history_handler:!0,__dummy__:"no comma"},EPS=1E-6; function invert(a){var b=a[0]*a[3]-a[1]*a[2];return[a[3]/b,-a[1]/b,-a[2]/b,a[0]/b,(a[2]*a[5]-a[3]*a[4])/b,(a[1]*a[4]-a[0]*a[5])/b]}function transform(a,b){return[a[0]*b[0]+a[2]*b[1]+a[4],a[1]*b[0]+a[3]*b[1]+a[5]]}function get_page_number(a){return parseInt(a.getAttribute("data-page-no"),16)}function disable_dragstart(a){for(var b=0,c=a.length;b=e.length||(e=e[a],e.loaded||this.pages_loading[a]))){var e=e.page,h=e.getAttribute("data-page-url");if(h){this.pages_loading[a]=!0;var d=e.getElementsByClassName(this.config.loading_indicator_cls)[0];"undefined"===typeof d&&(d=this.loading_indicator.cloneNode(!0), d.classList.add("active"),e.appendChild(d));var f=this,g=new XMLHttpRequest;g.open("GET",h,!0);g.onload=function(){if(200===g.status||0===g.status){var b=document.createElement("div");b.innerHTML=g.responseText;for(var d=null,b=b.childNodes,e=0,h=b.length;e=a?d.loaded?d.show():this.load_page(e):d.hide()}},update_page_idx:function(){var a=this.pages,b=a.length;if(!(2>b)){for(var c=this.container,e=c.scrollTop,c=e+c.clientHeight,h=-1,d=b,f=d-h;1=e?d=g:h=g;f=d-h}this.first_page_idx=d;for(var g=h=this.cur_page_idx,k=0;dc)break;f=(Math.min(c,l+f)-Math.max(e,l))/f;if(d===h&&Math.abs(f-1)<=EPS){g=h;break}f>k&&(k=f,g=d)}this.cur_page_idx=g}},schedule_render:function(a){if(void 0!==this.render_timer){if(!a)return;clearTimeout(this.render_timer)}var b=this;this.render_timer=setTimeout(function(){delete b.render_timer;b.render()},this.config.render_timeout)},register_key_handler:function(){var a=this;window.addEventListener("DOMMouseScroll",function(b){if(b.ctrlKey){b.preventDefault();var c=a.container, e=c.getBoundingClientRect(),c=[b.clientX-e.left-c.clientLeft,b.clientY-e.top-c.clientTop];a.rescale(Math.pow(a.config.scale_step,b.detail),!0,c)}},!1);window.addEventListener("keydown",function(b){var c=!1,e=b.ctrlKey||b.metaKey,h=b.altKey;switch(b.keyCode){case 61:case 107:case 187:e&&(a.rescale(1/a.config.scale_step,!0),c=!0);break;case 173:case 109:case 189:e&&(a.rescale(a.config.scale_step,!0),c=!0);break;case 48:e&&(a.rescale(0,!1),c=!0);break;case 33:h?a.scroll_to(a.cur_page_idx-1):a.container.scrollTop-= a.container.clientHeight;c=!0;break;case 34:h?a.scroll_to(a.cur_page_idx+1):a.container.scrollTop+=a.container.clientHeight;c=!0;break;case 35:a.container.scrollTop=a.container.scrollHeight;c=!0;break;case 36:a.container.scrollTop=0,c=!0}c&&b.preventDefault()},!1)},rescale:function(a,b,c){var e=this.scale;this.scale=a=0===a?1:b?e*a:a;c||(c=[0,0]);b=this.container;c[0]+=b.scrollLeft;c[1]+=b.scrollTop;for(var h=this.pages,d=h.length,f=this.first_page_idx;f= c[1])break}g=f-1;0>g&&(g=0);var g=h[g].page,k=g.clientWidth,f=g.clientHeight,l=g.offsetLeft+g.clientLeft,m=c[0]-l;0>m?m=0:m>k&&(m=k);k=g.offsetTop+g.clientTop;c=c[1]-k;0>c?c=0:c>f&&(c=f);for(f=0;fc.length;)c.push(null);var g=b||this.pages[this.cur_page_idx],d=g.view_position(),d=transform(g.ictm,[d[0],g.height()-d[1]]), g=this.scale,l=[0,0],m=!0,k=!1,n=this.scale;switch(c[1]){case "XYZ":l=[null===c[2]?d[0]:c[2]*n,null===c[3]?d[1]:c[3]*n];g=c[4];if(null===g||0===g)g=this.scale;k=!0;break;case "Fit":case "FitB":l=[0,0];k=!0;break;case "FitH":case "FitBH":l=[0,null===c[2]?d[1]:c[2]*n];k=!0;break;case "FitV":case "FitBV":l=[null===c[2]?d[0]:c[2]*n,0];k=!0;break;case "FitR":l=[c[2]*n,c[5]*n],m=!1,k=!0}if(k){this.rescale(g,!1);var p=this,c=function(a){l=transform(a.ctm,l);m&&(l[1]=a.height()-l[1]);p.scroll_to(f,l)};h.loaded? c(h):(this.load_page(f,void 0,c),this.scroll_to(f))}}}},scroll_to:function(a,b){var c=this.pages;if(!(0>a||a>=c.length)){c=c[a].view_position();void 0===b&&(b=[0,0]);var e=this.container;e.scrollLeft+=b[0]-c[0];e.scrollTop+=b[1]-c[1]}},get_current_view_hash:function(){var a=[],b=this.pages[this.cur_page_idx];a.push(b.num);a.push("XYZ");var c=b.view_position(),c=transform(b.ictm,[c[0],b.height()-c[1]]);a.push(c[0]/this.scale);a.push(c[1]/this.scale);a.push(this.scale);return JSON.stringify(a)}}; pdf2htmlEX.Viewer=Viewer;})(); try{ pdf2htmlEX.defaultViewer = new pdf2htmlEX.Viewer({}); }catch(e){}

A Personal Exposure Agent Based Model (ABM) with the capacity of aggregation at various

levels of population size

Dimitrios Chapizanis

1

, Spyros Karakitsios

1

, Dimosthenis Sarigiannis

1, 2, 3

(1) Aristotle University of Thessaloniki, Department of Chemical Engineering, Thessaloniki, Greece.

(2) HERACLES Research Center on the Exposome and Health, Center for Interdisciplinary Research and

Innovation, Thessaloniki-, Greece

(3) School for Advanced Study, Science, Technology and Society Department, Environmental Health

Engineering, Pavia, Italy

Innovations in sensors technology create possibilities to collect environmental and exposure-related data at

unprecedented depth and breadth. Measuring, though, personal exposure directly requires a large number of

people and therefore is often not feasible due to time and financial constraints. Considering the substantial

technical and ethical hurdles involved in collecting individual data for whole populations, this study introduces

a personal exposure model, where movement and interaction behaviour are simulated using Agent Based

Modelling (ABM), informed by sensors webs. The developed ABM allows us to quantitively assess personal

and community exposure differences and enables the identification of specific activities that may be linked to

higher levels of pollution. This approach permits the cost-effective construction of time-activity diaries and

daily exposure profiles, considering different microenvironments and socioeconomic characteristics. The

proposed method leads to a refined exposure assessment model that effectively addresses targeted subgroups of

population. It can serve as a tool to evaluate impacts of public health and environmental management policies

prior to implementation, reducing the time and cost required to identify effective measures.

ΑΒΜ is a simulation technique that allows us to explore and understand phenomena, where independent entities

interact together, forming an emergent whole. While the direct representation of individuals’ actions is

organisationally difficult, ΑΒΜ simplifies this process by managing information at the level of the autonomous

decision-makers, called “agents”. These heterogeneous actors have personal attributes and are programmed to

react and act in their environment while following a set of behavioural rules. By simulating actions and

interactions at the individual level, the diversity that exists among agents can be detected, as rise is given to the

behaviour of the system as a whole.

A city scale ABM was developed for urban Thessaloniki, Greece. Population statistics, road and buildings

networks data were transformed into human, road and building agents, respectively. In order to define

behavioural patterns for human agents, information retrieved from time-use surveys and literature associations

on how SES affects behavioural patterns was considered. The gathered evidence was implemented into the

ABM code in the form of behavioural rules. Rules are expressed as “if-then” statements or as functions that,

depending on personal sociodemographic attributes, define the probability of a human agent to proceed to a

certain action. Culturally varying personal attributes (such as age, gender, level of education) enter into these

functions, but the algebraic form remains fixed, as does the human agent’s practice of maximising the function.

An activity-selection function always assigns the next activity based on the Harmonised European Time Use

Survey (HETUS) dataset. Additional functions were also established; an interaction-function enables activities

sharing within human agents whereas a vehicle-selection function assigns the probability of choosing a specific

mode of transport when a virtual person is in transit. Individual characteristics (e.g. age, gender, income) provide

capabilities or constraints on the agents’ behavioural rules. In order to further inform and validate the model,

time-geography of exposure data, derived from a personal multi-sensors campaign on 150 households of urban

Thessaloniki, was used.

Overall, as a prevalence of an agent-specific decision-making and based on the distance between point of

departure and the targeted destination, virtual individuals of different sociodemographic backgrounds, use

different means of transportation and follow a different sequence and types of activities. Behaviours that were

not explicitly programmed into the model’s code, arise through the human agents’ interactions enabling the

examination of expected or unexpected emerging behaviours from the bottom-up. At the end of a model run,

spatiotemporal trajectories are coupled with spatially resolved pollution levels. Personal exposure, expressed as

inhalation adjusted exposure, was evaluated by assigning PM2.5 and PM10 concentrations to human agents

based on their coordinates, the type of location and intensity of the encountered activities.

The effect of sociodemographic characteristics in exposure refinement are clearly illustrated in the following

exposure profiles of two human agents living in the same neighbourhood (Figure 1).

Figure 1. Time-Activity Information and PM2.5 exposure profiles of 2 human agents living in the same neighbourhood

Even though these virtual people they are both phenomenally exposed to similar outdoor air pollution levels

(PM2.5 daily intake dose of 7.8 mg/kg-day), their inhalation adjusted exposure (PM2.5 daily intake dose of 5.8

and 8.3 mg/kg-day respectively) varies significantly (53% change rate) due to the prevalence of different

behaviours. This methodology proves that if calculations were only based on outdoor modelled - or retrieved

by the nearest Air Quality station - residential PM concentrations, as commonly established in current studies,

then personal exposure could be significantly mis-estimated. Crucial exposure differences would not have been

identified, having disregarded the time one spends in different microenvironments and the intensity of enrolled

activities. ABM findings indicated that inhalation adjusted exposure differences between 2 individuals living in

the same neighbourhood can vary by as much as 87%, due to radically different spatiotemporal behaviours. This

major difference was observed for cases where one human agent has a full-time job and often exercises in

outdoor courts whereas his/her neighbour is a retired homemaker.

Grouping individual exposure profiles based on criteria such as age, gender, SES indicators or based on the area

of residence/work/study, also enables the extraction of representative exposure profiles for subgroups of

population. Indicative ABM results capture PM2.5 exposure differences between different communities of

major interest (Figure 2) and indicate that the four most exposed groups are: child agents, adult male agents

with a full-time job, low-income male agents residing in western Thessaloniki region and human agents of a

lower educational level. Subgroup median exposure for these cases was notably higher than the total population

median value.

Figure 2. PM2.5 daily intake dose variation for different groups of the total human agents population

The validity of the ABM results was examined by performing a compatibility check between the ABM-retrieved

virtual exposure data and the real exposure profiles extracted from the sensors campaign. Findings indicated

that the difference between the median values in box plots of the same subgroup of (virtual vs. real) population

does not exceed the range of 5-15%. This serves as a first demonstrates that the ABM-derived emergent

trajectories are compatible to reap-world spatiotemporal behaviours, which is of crucial importance since space-

time and activity information is a key determinant of personal exposure.

By modelling the heterogeneous routine of human agents, the developed ABM can produce detailed information

related to the societal system examined and can generate data that could be used to fill in the gaps that exist in

traditional exposure-related datasets. This method signifies a step forward, away from the earlier static nature

approaches of urban modelling, where total population is divided into homogeneous subpopulations. The

establishment of an ABM approach that integrates Socio-Economic Status (SES) indicators with the capacity

for aggregation and analysis at various levels of population size, leads to an exposure assessment model

especially useful for vulnerable groups of population, such as children, the elderly and people living in hot spot

areas. This is an opportunity to capture evidence for cases where specific subgroups of population are

disproportionately exposed to higher levels of environmental risk than other parts of the same society. This is,

therefore, a method where cases that exhibit environmental injustice can be detected and interpreted through an

artificial society type model. Moreover, the proposed method can be used for evaluating the probable impacts

of different public health policies prior to implementation reducing, therefore, the time and expense required to

identify efficient measures.

Keywords: personal exposure assessment, agent based modelling, sensors technology