Tentamen (uitwerkingen)
tededede
Tentamen van 1 pagina's voor het vak Financial Management BDK aan de RuG (dtette asdfa)
[Meer zien]Voorbeeld 1 van de 1 pagina's
Voorbeeld 1 van de 1 pagina's
In winkelwagengesponsord bericht van onze partner
Verkoper
Volgen
Voorbeeld van de inhoud
Minds for Mobile AgentsAndrew Heathcote1,2 (andrew.heathcote@newcastle.edu.au)
Charlotte Tanis2
Jonne Zomerdijk Tessa Blanken Dora Matzke Denny Borsboom 2
2 2 2
1
School of Psychology, University of Newcastle 2Department of Psychology, University of Amsterdam
Objectives Strategic Model
Effective training through simulation of multi-agent interactions
25
Figure 2. Initial configuration
requires autonomous agents guided by realistically complex with an agent entering. “+”
20
goals and mental models. We develop a framework that can symbols are waypoints and “o”
build, parametrize, and simulate systems of such agents. symbols are the agent’s
15
shopping goals. Simulation
studies conducted by assigning
Background
y
each agent a randomly
10
Although there is a rich literature on the measurement and modeling of selected set of goals to satisfy
showed the operational model
human movements in dense crowds with simple goals (e.g., moving to
5
was competent at satisfying
an exit, e.g., Helbing et al., 2000) the associated “people as goals while avoiding collisions
0
and avoiding gridlock even at
A
homogenous particles” approach is questionable in more common
high densities. 0 10 20 30 40
scenarios where agents with diverse characteristics follow individual x
Each agent has a sequence of of spatially defined goals of two types:
plans requiring navigation through a complicated series of goals. 1) “must visit” goals satisfied when moving within a threshold distance
We instantiate the same operational level functionality (relatively 2) “way-point” goals satisfied when the following goal becomes visible
automatic abilities enabling movement towards goals while avoiding
Route-finding and path optimization (i.e., “traveling salesman”)
obstacles) as social-force models (Campanella et al., 2014) in a utility- algorithms can be used to initially build sequences (“goal stacks”) that
maximizing discrete-choice framework that models step decisions satisfy sets of must-visit goals and other constraints such as one-way
(Robin et al., 2009). We also add strategic level route finding (Larmet, regions, and random perturbation used to mimic sub-optimal solutions.
2019) and path planning (Hahsler & Hornik, 2007) abilities, allowing
agents to plan and re-plan individualized series of goals. Operational factors compromising path plans (e.g., being pushed off
course by other pedestrians) can be addressed by re-planning using
Operational Model the same methods or by path re-tracing to the last point where the next
goal was visible. Re-planning can also be used with changing
conditions (e.g., crowds blocking the planned path at high densities).
Figure 3. Path plan satisfying 5
25
Path = 86 m
> must-visit goals (LETTERS)
Random utility (U) guides k=33 (11 direction “cones” x 3 velocity with one-way constraints (grey
20
> < arrows). Naturalistic paths
“rings”: slow/constant/speed up) step choices (rU = randomness). x
< > were obtained by a simple
15
$% &' (' )* +' ,- .+ greedy algorithm and allowing
𝑈!"# 𝑈!" + 𝑈!" + 𝑈!" + 𝑈!" + 𝑈!" + 𝑈!" + 𝑈!" ⁄𝑟/
C
y
way-point goals to be satisfied
10
before the agent gets close to
Utility is a sum of scaled (b) power functions (exponent a) of absolute D
them (e.g., in the orange path
differences (d≥0) between the current position (meters) or velocity < < the two way-point goals will be
5
E B A
satisfied early to produce a
(i.e., speed, m/sec, & angle, degrees/90) and avoidance or approach smoother trajectory around the
goals. Scoring disutility = - utility:
0
x
end of the aisle).
0 10 20 30 40
Repulsion Disutility = b/da Attraction Disutility = bda x
Conclusions and Future Directions
Individual Attraction Utilities
Step-choices based on utility provide a flexible framework to
PS = Individuals preferred speed (threshold linear slowing near goals)
provide minds for mobile agents, allowing them to operate
GA = Goal angle (tendency to head to goals) competently and independently in dynamic multi-agent
CA = Current angle (tendency to continue ahead, with different weights environments.
for left and right sides to account for side preference)
The discrete-choice framework enables likelihood-based fitting,
Social Repulsion Utilities providing parameter estimates that illuminate the psychological factors
ID = Interpersonal distance (minus infinity at body overlap) changing behavior in different contexts.
BA = Blocked angle (avoid cones with lots of pedestrians) Goal stacks augmented with re-planning provide a flexible means of
Social Attraction Utilities instantiating dynamic control of complex spatial navigation plans.
FL = Follow the leader (promotes lane forming in crowded situations) The model is being used in projects investigating the effects of space
WB = Walk beside (social group dependent, also biases FL) design and movement rules on social distancing and virus spread.
These 7 components produce plausible behavior in complex scenarios Future work will use position data from movement experiments to
(e.g., shopping in a supermarket), and the framework can be easily quantify individual differences, enabling the model to be calibrated for
extended to add on new behaviors appropriate for specific contexts veridical simulations of complex real-world scenarios.
(e.g., avoiding visibility to other agents). The same framework is being used to develop cognitive models of
strategic decision making in a navel escort simulation task requiring
participants to protect a high value target from a submarine.
25
p o
References
L
g
R A
20
m
f M k
N d n
Campanella, M., Hoogendoorn, S., & Daamen, W. (2014). The Nomad model: Theory,
O
a l developments and applications. Transportation Research Procedia, 2, 462–467.
I P
u s Hahsler M, & Hornik K. (2007). TSP - Infrastructure for the traveling salesperson problem. Journal
15
h r
H
of Statistical Software, 23, 1-21.
y
B
D
Q
Helbing, D., Farkas, I., & Vicsek, T. (2000). Simulating dynamical features of escape panic. Nature,
407, 487–490.
10
e t E
C
v Larmet V (2019). “cppRouting: Fast implementation of Dijkstra algorithm in R.
w
U q https://github.com/vlarmet/cppRouting>.
5
b T
Robin, T., Antonini, G., Bierlaire, M., & Cruz, J. (2009). Specification, estimation and validation of a
K
G c
F
j S pedestrian walking behavior model. Transportation Research Part B, 43, 36–56.
Acknowledgements
0
0 10 20 30 40
x
Figure 1. Agents moving around a 40x25m supermarket (black rectangles are shelves). Red circle
= 0.6m diameter body; blue circle = 1.5m diameter; red/black arrow = goal/current direction
(disparities between arrows occur where pedestrians are avoiding each other); grey body = pause https://dataversuscorona.com/
at goal. Agents enter on the bottom left and exit through one of two exits on the middle/top left. https://www.dst.defence.gov.au/partner-with-us/university/rn-uds
R E S E A R C H P O S T E R P R E S E N T A T IO N T E M P L A T E © 2 0 1 9
www.PosterPresentations.com
Voordelen van het kopen van samenvattingen bij Stuvia op een rij:
Verzekerd van kwaliteit door reviews
Stuvia-klanten hebben meer dan 700.000 samenvattingen beoordeeld. Zo weet je zeker dat je de beste documenten koopt!
Snel en makkelijk kopen
Je betaalt supersnel en eenmalig met iDeal, creditcard of Stuvia-tegoed voor de samenvatting. Zonder lidmaatschap.
Focus op de essentie
Samenvattingen worden geschreven voor en door anderen. Daarom zijn de samenvattingen altijd betrouwbaar en actueel. Zo kom je snel tot de kern!
Veelgestelde vragen
Wat krijg ik als ik dit document koop?
Je krijgt een PDF, die direct beschikbaar is na je aankoop. Het gekochte document is altijd, overal en oneindig toegankelijk via je profiel.
Tevredenheidsgarantie: hoe werkt dat?
Onze tevredenheidsgarantie zorgt ervoor dat je altijd een studiedocument vindt dat goed bij je past. Je vult een formulier in en onze klantenservice regelt de rest.
Van wie koop ik deze samenvatting?
Stuvia is een marktplaats, je koop dit document dus niet van ons, maar van verkoper Jonnez. Stuvia faciliteert de betaling aan de verkoper.
Zit ik meteen vast aan een abonnement?
Nee, je koopt alleen deze samenvatting voor €7,49. Je zit daarna nergens aan vast.
Is Stuvia te vertrouwen?
4,6 sterren op Google & Trustpilot (+1000 reviews)
Afgelopen 30 dagen zijn er 50155 samenvattingen verkocht
Opgericht in 2010, al 14 jaar dé plek om samenvattingen te kopen