21
Automata

Automata

  • Upload
    imala

  • View
    175

  • Download
    0

Embed Size (px)

DESCRIPTION

Automata. States and transitions. Bathe. Get dressed. Sit. Get dressed and sit. Sleep. Wake up. Variant 1. Bathe. Get dressed by Mom. Get dressed by Dad. Sit. Get dressed and sit. Sleep. Wake up. Variant 2. Bathe. Get dressed. Sit. Get dressed. Sleep. - PowerPoint PPT Presentation

Citation preview

Page 1: Automata

Automata

Page 2: Automata

States and transitions

Get dressed

Sit

Get dressed and sit

Sleep

Bathe

Wakeup

Page 3: Automata

Variant 1

Get dressed by Mom

Sit

Sleep

Bathe

Get dressed by Dad

Get dressed and sit

Wakeup

Page 4: Automata

Variant 2

Get dressed

Sit

Sleep

Bathe

Get dressed

Page 5: Automata

States as combinations

and

Activity

Mood

Page 6: Automata

Applications

Parsing (for compilers or other software)

Verifying program correctness (model-checking)

Modeling anything

Page 7: Automata
Page 8: Automata

Stages/States

SSG: Self-sufficiency in growth signals

IAG: Insensitivity to anti-growth signals

Ang: Sustained angiogenesis

LRP: Limitless replicative potential

Ev Ap: Evading Apoptosis

M: Metastasis

Page 9: Automata

A Possible Drug Therapy

VEGF signals for creation of new blood vessels

The drug Avastin inhibits the VEGF signaling pathway

This prevents tumors from getting an adequate blood supply

Page 10: Automata

Using Avastin

Normal

Avastin inhibits VEGF pathway

Page 11: Automata
Page 12: Automata

Evaluating therapies

Cost? Pareto optimality

Other properties? Model-checking:

Metastasis is never reached Following angiogenesis, evading apoptosis is

never acquired

Page 13: Automata

Model shortcomings

Ignores time In pancreatic cancer, about a year for K-ras

mutations to lead to abnormal growth of tissue

Seventeen years for a large benign tumor to evolve into an advanced cancer

Two years within the advanced cancer for cells to metastasize

Drugs completely switch off transitions

Page 14: Automata

Adding time

Use a collection of clocks.

Clock rates may be different and can change independently!!

Time measured by heartbeats

Time measured by bars

Time measured by turns

Page 15: Automata

Adding time

Time is measured by a collection of clocks

State invariants: Maximum time on each clock in a given state

Transition constraints: Time on each clock in origin state before transition can happen

Drug effects: How the drug affects clocks

Page 16: Automata

Timed Automata

Page 17: Automata

Time invariants for transitions

Trans: Scrub ≥15

Trans:Scrub≥10

Trans: Head turns ≥60

Trans:t≥30

Trans: t≥120

Trans: Breaths≥1000

Page 18: Automata

Time constraints for states

Scrub≥30

Head turns≥100

t≥120

Breaths≥2000

Page 19: Automata

Drug effects

Scrub≥30

Head turns≥100

t≥120

Breaths≥2000

Trans: Scrub ≥15

Trans:Scrub≥10

Trans: Head turns ≥60

Trans:t≥30

Trans: t≥120

Trans: Breaths≥1000

Oatmeal: reduce head turns by 50%, breaths by 10%Milk: reduce head turns by 75%, breaths by 20%Sugar: double head turns and breathsValium (administered to Mom): reduce Scrub by 50%

Page 20: Automata
Page 21: Automata

CHA Therapy

IAGSSG Ang4 2

4

Can only stay in state until clock A reaches 4

Can’t leave state until clock B reaches 2Avastin: slows clock B by .5