Software design for Smart Logistics - Universiteit Twente · Software design for Smart Logistics....

Post on 30-Apr-2020

10 views 0 download

Transcript of Software design for Smart Logistics - Universiteit Twente · Software design for Smart Logistics....

Lunch talk BMS Smart Industry Research Roadmap

05-06-2018

Software design for Smart Logistics

Current role and ambition:• Business Development Manager at CAPE Groep• Prospective industrial PhD candidate University Twente

Professional experience:• 10 years industry experience in supply chain, logistics and IT• Member of TKI Dinalog supervisory board 4C program 2020• Involved in multiple topsector projects: 4C & Synchromodal Logistics• Part-time lecturer at University Twente, Saxion UAS, NHTV UAS• Initiator of the vLm round table Twente

Educational background:• Master International Supply Chain Management• Bachelor Business Engineering• Bachelor Human Resource Management

Sebastian Piest

Partnerships

Agility

Innovation

Relations

TMS

WMS

APS / Maps

Finance

Salaris

Back-officeiPaaS

Customs

Re

ceiv

er

Proof of Delivery (POD)

Track & Trace / ETA

Dispatching & Labeling

Shipment planning

Ass

ets

/ L

SPs

/ C

arri

ers

Cu

sto

me

r

Orders (EDI / Web)

Quotes (km / € / CO2)

aPaaS

ERP

FMS

Etc.

Data analyses / BI

Reports / Dashboards

Message exchange

Notifications & Alerts

Invoicing & self billing

B2B / Devices

8

Triple helix projects; making science actionable

• Based on Artificial Intelligence (AI) and the Intelligent Amplification (IA):

– IA is first proposed in 1950s and 1960s, but contrasted by AI developments

– IA uses information technology in augmenting human intelligence

– IA proposes a symbiotic partnership between human and AI

– Combining AI and IA utilizes strengths of both concepts

• “Human-in-the-loop” concept:

– Human defines strategy, oversees AI, corrects when needed

– AI executes routine tasks according to strategy

• Fly-by-wire example

Human in the loop concepts for decision making in logistics

Source: A. Dobrkovic (2016)

• Identify suitable AI environment:– Repetitive tasks

– High workload

– Limited external knowledge required

– Time limits

• Access soft knowledge via IA:– Information available to human

decision maker

– Not every case can be included in the model, exceptions will keep occurring

– Creativity is essential for making the “right” decisions on complex problems

AI/IA in operational business environments

Source: A. Dobrkovic (2016)

Framework for Enhancing Scheduling Processes

Source: A. Dobrkovic (2016)

Learning / personalizing optimization policies

Scenario Input

States

TacticsThresholdsTolerances

PolicyGoals

data mining

Domain expert

Source: A. Dobrkovic (2016)

Execution flow

PolicyGoals

Scenario Input

Intelligentagent

Consult

Execute

Transfer control

low conf.

Source: A. Dobrkovic (2016)

Level 1: top level processes

Level 4: decomposed process elements

Level 3: decomposed processes

Level 2: process categories

RFI RFQAnalysis &

negotiationContracting

PO Management

Container trucking

Customs services

Freight insurance

Carrier selection

Freight allocation

Sea- Air FreightTransportation Warehousing Distribution

Ocean carrier websites Sailing schedules and departure times

Marine TrafficReal-time actuals on live map

Terminal websites Terminal planning and arrival times

Purchase order statusMonitor the actual- and estimated departure and arrival times

• Increase productivity of the workforce:– Automate manual routine tasks with AI– Extend professional judgement & experience with IA

• Change management:– Stimulate the willingness to share individual knowledge– Bundle individual tactics into a shared policy and goals– Learn to trust and work with data captured by agents– Feed the system and continously refine data quality– Change the way of work

• Integration with 3rd party websites:– No API to connect with– Some require user log-in / commercial subscription– Have their own structure / use different input values

Synchromodal IT goal and challenges

Web

site scrapin

g

Integratio

n layer

Visib

ility layer

Back-office

NLIP Services

iSHARE

Real-time ETA monitoring of perishables

Daily operations - perishables

Real-time:

ETD/ATD

ETA/ATA

Webscraper robots

Real-time, open data

Integration of ETA Monitor and webscraper

ETA Monitoring Dashboard

• AI will dramatically change the way we work, but is not a “silver bullet”

• Routine tasks in daily operations and decision making can be automized by AI

• Soft knowlegde of domain experts can be accessed via IA to optimize policy goals

• The “Human-in-the-loop” concept combines the best of both approaches:– Human defines strategy, oversees AI, corrects when needed

– AI executes routine tasks according to strategy

• At the end of the day, complex problems and exceptions still require human intelligence

Lessons learned from Synchromodal IT

Current research projects

EMAGIZOrchestration, integration & monitoring

SMEUSERS

MENDIXSaaS basedSwiss Knife

solution withAPIs

ALGORITHMSAPI enabled

micro- servicesMINERS

User interface CLDM Connected algorithmsAction based learning

“Autonomous Logistics Miners for SMEs”

DataRel: Big Data for Resilient Logistics

Smart pallet pooling / VMI

Real-time traceability in food retail

Ahrma

Cloud

Place totes

on roll

containers

Load trucks

with roll

containers

Deliver roll

containers

to X-Dock DC

Return roll containers

IoT

data

Reporting /

Dashboards /

Portal(s)

Order

data

& IoT

data

Software design for Smart Logistics

Lunch talk BMS Smart Industry Research Roadmap

05-06-2018

Thank you for your attention