1

SFB 637

International Conference on Dynamics in Logistics,August 28th - 30th, 2007, Bremen, Germany

Dynamic Decision Making on Embedded Platforms in Transport Logistics –

A case study

Reiner Jedermann, Luis Javier Antûnez Congil, Martin Lorenz,Jan D. Gehrke, Walter Lang and Otthein Herzog

Institute for Microsensors, -Actuators and -Systems IMSAS

Center for Computing Technologies – TZI

IM SASIM SAS

2

SFB 637

Outline

Introduction

Decentralized route planning

Agent-based shelf life supervision

Extension for multi package problem

Experimental evaluation

Summary and future work

3

SFB 637

Introduction

Shifting intelligence from central control to transport containers

Complexity and cost pressure in supply chains forces new approaches

Individual planning for each palette / freight item- Transportation of perishable goods- Setting with high amount of data per cargo for monitoring- Unexpected changes in product quality may force re-planning

(change of vehicle and/or destination)- Vision: intelligent cargo- Current hardware solution on vehicle/container level

Two points of view- Planning for full truckloads (existing demonstrator)- Combined planning for part loads (simulation of new concept)

4

SFB 637

Bremen

Berlin

Agent fortropical fruits

Recommended temperature overstepped.

Shelf life reduced to 24 hours.

Currently planned route

requires 36 hours

Search for alternatives / rescue plan

Autonomous transport supervision

Individual software agents to supervise each freight item

Adapts to individual requirements of the loaded goods

Asses the influence of deviations of the environmental parameters (temperature) to the freight quality

Triggers re-planning if some risk is detected

Current solution: freight quality evaluation within vehicle/container

Vision: Intelligent Cargo

PlanningAgent

5

SFB 637

Autonomous Transport Scenario

6

SFB 637

Route Planning

Autonomous routing for perishable goods has to consider shelf life criteria and dynamic environments, e.g., (unexpected) quality changes

Routing problems: TSP, VRP, VRPPD, VRPTW

(Optimal) routing solutions are NP-hard in general

- Constrains dimensions of maximum problem space

- Limits practicability for embedded systems

Cost function with shelf life is subject to information privacy concerns when using external routing services

Heuristic (sub-optimal), cooperative (distributed) approaches needed

7

SFB 637

Software Agents

What are agents?

‘‘An agent is a computer system situated in some environment, and that is capable of autonomous action in this environment in order to meet its design objectives” (Jennings & Wooldridge 1998)

Autonomous agents act without direct intervention of others

Multiagent systems: agents communicate and cooperate to solve complex problems that are beyond the capability of a single agent

Pro

pose

Accept/R

eject

A1

A2

8

SFB 637

Software Agents

What are agents?

‘‘An agent is a computer system situated in some environment, and that is capable of autonomous action in this environment in order to meet its design objectives” (Jennings & Wooldridge 1998)

Autonomous agents act without direct intervention of others

Multiagent systems: agents communicate and cooperate to solve complex problems that are beyond the capability of a single agent

Multiagent system architecture and communication is standardized by the IEEE Foundation for Intelligent Physical Agents (FIPA)

FIPA multiagent runtime environments: JADE and LEAP

Agent architectures: e.g. BDI, goal-oriented architecture, with autonomous goal selection (deliberation), and means-end reasoning

Pro

pose

Accept/R

eject

A1

A2

9

SFB 637

Autonomy in Software Agents

Levels of Autonomy (Timm 2006) Strong regulation: No autonomous capabilities; every decision is

determined by external entities (reflex agent architectures).

Operational autonomy: Competence to choose course of action in predefined strategic boundaries(goal-oriented architectures, means-end reasoning).

Tactical autonomy: Enables the system to deliberate on different alternatives for operational behavior (BDI architectures, deliberation).

Strategic autonomy: Conventionally determined by the system designer (desires and algorithms). Beyond classic BDI architecture.

Autonomy in Case Study Local vehicle agent has operational autonomy: route selection

Possibly tactical autonomy: customer/cargo preference adaptation

10

SFB 637

Example shelf life (Lettuce)

0

2

4

6

8

10

0 5 10 15 20

Temperature °C

Sh

elf

life

/ lo

ss in

day

s .

Shelf life(T)

Loss per Day

Reference temperature 6 °C

Tripple speed of quality decay at 14 °C

Activation energy for Lettuce

1

11

SFB 637

Local processing

Sensor raw data

0

6

12

18

0 10 20 30

T (°C)

KQ

(d

ay

s)

TransportOperator

QualityModelling

StandardT&T

0

6

12

18

0 10 20 30

T (°C)

KQ

(d

ay

s)

TransportOperator

QualityModelling

Standard T&T + Processor

QualityInformation

12

SFB 637

Implementation of agents on embedded systems

Software agents on embedded systems ARM Processor 1 Watt @ 400 MHz Embedded real time JAVA runtime environment Implementation of a reduced agent platform (JADE –

LEAP)

Adapt agent systems to the restrictions of embedded systems: Power, memory and computation resources

100%

< 5%

~0.1%<< WSN

13

SFB 637

Hardware

RFIDReader

Local Pre-Processing

SensorNodes

ExternalCommu-nication

FreightObject(RFID)

14

SFB 637

Agent Transmission process

Intelligent Agent

Transport- and handling- instruction Supervision in behalf of the owner

Logistical object Passive RFID-Label

Intelligent truck or container

CPU platform Sensors RFID-Reader

Reading RFIDat loading

Truck requests agent at loading

Dynamic link

RFID

15

SFB 637

Evaluation of sensor data

0

6

12

18

0 10 20 30

T (°C)Sh

elf

Lif

e (

da

ys

)

Humidity

Quality-Index

Temperature

16

SFB 637

Multiple package problem

Setting: Truck contains several pallets of perishable goods for different destinations. In which order should the destinations be served to deliver the goods before

expiration?

Additional requirements High local temperature deviations force individual supervision Simply multiplying the number of agents also multiplies the amount of

communication Truck / Container has to find a route that serves the individual needs of the

majority of all loaded packages

Planned improved solution Extension towards the current demonstrator software of intelligent container Idea: Reducing communication by shifting part of the route planning into the

means of transport Simulation Further improvement by increasing the level of autonomy

17

SFB 637

1.5

2

2.5

3

0

0.5

1

1.5

2

0

0.5

1

1.5

2

4.87°C

4.01°C

5.06°C

3.97°C

3.33°C

5.11°C3.08°C

1.71°C

2.60°C

1.79°C

1.40°C

2.64°C

0.70°C

1.77°C

0.49°C

Temperature along the xyz-axis

Average of reefer side ~2 °C colder than other side

Single loggers behave 'chaotic'

18

SFB 637

The test case

Extension of the Traveling Salesman Problem Not shortest way, but minimize shelf life losses by route planning Dynamic form: unexpected changes of shelf life and traffic jams

Item Nr

Desti-nation

Initial

Shelf life

1 Town 7 12 hours

2 Town 3 50 hours

3 Town 1 36 hours

… … …

… … …

Distance Town

1

Town

2

Town

3

…

Town 1 - 5 hours 7 hours …

Town 2 5 hours - 3 hours …

Town 3 7 hours 3 hours - …

… … … … …

19

SFB 637

Distributed Planning by truck agents

Route Planning Agent (RPA) Remote Server Access to road maps and traffic

information Public information

Local Vehicle Agent (LVA) Embedded System (Truck) Evaluates Shelf life Private information

Goal fulfillment Maximize sum of remaining shelf life

at delivery Strongly avoid zero shelf life /

expired products

Request route proposals

Set of suggestions with low driving time

20

SFB 637

Experimental evaluation

Distributed heuristic solution Software simulation Comparison with optimal solution

Process repeated in each town Unit: Travel distance in hours

0 2 4 6 8 10 12 14 16-1

0

1

2

3

4

5

6

7

0*

1*

2*

3*

4*

5*

6*

7*

8*

9*

10*

11*

First planning step

Shortest round trip through remaining towns

Some decision alternatives

Starting Point

Current Position

Continue counter

clock wise

Continue clock wise

21

SFB 637

Experimental evaluation 2

Replanning Change of planned route in step 2 caused by new information Caused by new route suggestions or Changed shelf life / traffic situation

0 2 4 6 8 10 12 14 16-1

0

1

2

3

4

5

6

7

0*

1*

2*

3*

4*

5*

6*

7*

8*

9*

10*

11*

First Suggestion (84.6 points)

Final Route (269.2 points)

Starting Point

Re-Planning

22

SFB 637

Experimental evaluation 3

Comparison to optimal solution In most cases solution close to optimum But hard to find if big difference between short route and optimal solution

0 2 4 6 8 10 12 14 16-1

0

1

2

3

4

5

6

7

0*

1*

2*

3*

4*

5*

6*

7*

8*

9*

10*

11*

Heuristic Solution (269.2 points)

Optimal Route (276.1 points)

23

SFB 637

Results

Summary of experimental results 600 “runs” with identical town-map and random initial shelf life values The points give a measure for the remaining shelf life at delivery. In 2/3 of all experiments the same number of packages had sufficient

remaining shelf life at delivery as in optimal solution (Row A) In average the remaining shelf life was 92% of the optimal possible value In the remaining 1/3 of experiments more packages as in the optimal solution

had zero shelf life (“lost packages”) at delivery (row B)

Runs Local planning Optimal Ratio

A (no losses) 402 252,73 points 272,02 points 92,62% ± 7,37

B (with losses) 198 More package losses as optimal solution

24

SFB 637

Summary and Future work

Case study for an autonomous logistic process Reduced

communication costs Lower computation

resources needed Continue locally if

communication fails Privacy Higher degree of

autonomy by enhanced architecture to change strategy if required (replacing software components on request)

Simple HeuristicLVA

Evaluation of Solution

Acceptable?LVA

Change of planning strategy

Time window based route optimization

RPA

No

25

SFB 637

The End

Thanks for your attentionwww.intelligentcontainer.com

ContactDipl.-Ing. Reiner JedermannUniversität Bremen, FB1 (IMSAS), Otto-Hahn-Allee NW1, D-28359 Bremen, GermanyPhone +49 421 218 4908, Fax +49 421 218 4774 rjedermann@imsas.uni-bremen.de

Dipl.-Inf. Jan D. GehrkeUniversität Bremen, FB3 (TZI), Am Fallturm 1D-28359 Bremen, GermanyPhone +49 421 218 8614 , Fax +49 421 218 7196jgehrke@tzi.de