PLANNING THE PROJECT MANAGEMENT WAY: EFFICIENT PLANNING BY EFFECTIVE INTEGRATION OF CAUSAL AND RESOURCE REASONING IN REALPLAN pdf

TECHNOLOGY DEVELOPMENTSTRATEGY $$$ MANAGEMENT MS Project Scheduling Tool SCHEDULE TASKS MID/ LOW MANAGER FEASIBLE PROJECTS BUDGETED, APPROVED PROJECT PLAN... YES Planning Done PLANNER Se

TECHNOLOGY DEVELOPMENT

STRATEGY

$$$

MANAGEMENT

MS Project Scheduling Tool

SCHEDULE

TASKS MID/ LOW

MANAGER FEASIBLE PROJECTS

BUDGETED, APPROVED PROJECT PLAN

0.0 20.0 40.0 60.0 80.0 100.0

#robots

0.0 0.1 1.0 10.0 100.0 1000.0

Performance of different Planners

in Shuffle 6 Blocks Problem

FF AltAlt RealPlan Graphplan Blackbox (Satz)

on_B_Tab

on_B_Tab on_A_Tab

PICK_A_R1 PICK_A_R2 PICK_B_R1 PICK_B_R2

.

on_A_B

Fact level 0

Action level 0

Fact level 1

Action level 1

Fact level 2 STACK_A_B_R2

1 3 5 7 9

# robots

0.1 1.0 10.0 100.0

Performance of Graphplan and Blackbox (satz)

in the 6-block Shuffle Problem in Blocks World

GP-G GP-S GP-TOT BB-TOT

1 3 5 7 9

# robots

0.1 1.0 10.0 100.0 1000.0

Performance of Graphplan in Shuffle Problem

with varying number of blocks and robots

SHUF4-GP-TOT SHUF6-GP-TOT SHUF8-GP-TOT SHUF10-GP-TOT

1.0 3.0 5.0 7.0

# robots

10.0 12.0 14.0 16.0 18.0

A Look into Plans by Graphplan

in the Blocks World domain

Length of plan Number of steps in plan

Ք̼ÉZÖ©Ï§Í¨È Ë¼Ó§ÈÍ Iê Iå ˼ӧÈÍ Iê I[å

A C

YES Planning Done PLANNER

Set Alloc Policy

SCHEDULER

Schedule Done YES

Post-process

YES Executable Plan NO

NO

YES Alloc needed

SCHEDULER

T A S L A T O R

TRANSLATED FEEDBACK

Plan with 2 resource

Level 1 Level 2 Level 3

Level K Level K-1

.

Maximally parallel plan

Maximally serial plan

Plan with R-1 resource Plan with R resource Plan with N >= R resource

à ӾԌÞwÈÈ Þ>“ ÚŒÍ¨Ç Ë¼ÚŒÍ€Ì¼ÍΠ̼ÍÍ¨ÇŒÉˆÛŒÔˆÚ Ì¼Í¨Ê†ÉˆÛ”Ì—Ù‰Í¨Ê"ˆÉ1É)éZ̗̀ىɈÈÍÎZÇjÝî̼ͨʆɈ۔̗ىÍىɈÇg”©Ó§Ù‰Ë€×aɈǔÍ

Level Actions by level # Robots

Unstack_R_blkF_blkE 1 1

5 6 6 7 8 9

Unstack_R_blkE_blkD Unstack_R_blkD_blkC

Unstack_R_blkB_blkA Stack_R_blkF_blkC Pickup_R_blkA Stack_R_blkB_blkF Stack_R_blkE_blkB Stack_R_blkA_blkE Stack_R_blkD_blkA

5 Putdown_R_blkC

Unstack_R_blkC_blkB 4

3 2

10

3 4

5 5 4 3 2 1 2

à ӾԌÞ(È[Þgê ̼ͨʼɈی̗ى͉ÐLΠ֩ʆˆ̤ÎZىˆͨÒîʆɈϧ۔˼ӾɈÇ*æ½ÉZÌx—‹V‘ •.Ք̼ÉZ֩ϾͨÈÞíöی̼éZͨÒϧӧǔͨʭʗڌÉ0Ñ Ì—Í‰Ð

A^2_5 A^2_6

Viewing actions as tasks of unit time

à ӾԌÞlȒâ[ÞX]Ó¾Í€Ñ ÉZæ˼ڌÍ(̼ͨʆɈی̤ى͉ÐLΠ֩ʆˆ̗ÎZىˆͨÒ1թϧÎZÇ1Ó§Çà ӾԈی̼͜ÈxÎZÊ­Îx˼ÎZʆänj̀˙Ñ@ÉZ̗äGˆɝֹÍ

ÎVø ø ÷

Ü

ø<ÿÜJÞ

ÿ 

ÜJÞù*ú

)'L&K„mV$U^IZP%mx<Vm|V%V$KŠG&V%i5V$GdO Ç I5KW`Íel¯¤oV$U"V%YRbU"V5k

LdmO‡V$K„m^LdUVW]U^IZmL&RbKlk)O"TpI5K

)'P\I5K–YU"V$V[Ldm‡IZY¶m^V%U‡G&VibV%G@͗ORqO"TpI5K

”('ÑP\I5Kˆ]O"V[Ldmgm^R

Classes of Resource Allocation Scheduling

EASY

HARD

[3 4] FIX [2] SAMELEN

R5YqITU"R)P%V$W]UI5GUV$O"Rb]U"P$VŠI5G&GdR)P$IZm^LdRbK I5G&M5RbUL!m^oS L&OˆOoR7_K L&K ±…L&Mb]UV‘³Ãe¤¯¤oL&O

)'ÑI5UVHU"V$G&ViZI5K„m…Irm8,

e±pRbU@V$I5P0ot]KpI5G&GdR)P$IZm^V%W

UV%G&V%iZI5K„mCOTpIZKk7Ldm(P0oV%P0~“OwmR,O"V$VgLdY“m^oV%UV|LdOI_|I\‰mRžIZOOLdMbKœI,UV%ORb]UP$VZeZ¯¤oVgP0oV%P0~œLdO

N^D_N^`dJdabL

$ cJMafe

A^1_6 A_7 A_8 A_9 A_10 A^1_5

A_4 A_3 A_2 A_1

A^2_5 A^2_6

Viewing actions as tasks of unit time (5 or more robots)

1

2 3

A^1_6 A_7 A_8 A_9 A_10 A^1_5

A_4 A_3 A_2 A_1

A^2_5 A^2_6

Viewing actions as tasks of unit time (3-4 robots)

1

2 3

1

2

2 3

ă ӾԌÞHăbỈ Þ ï ٗڌͨҩ۔ϧӧ̌ÔîỊ¼ÎZ͆ố̌̀Ị™Ñ@ĨZ̼ô ĨZư—‹j‘P$•Ք̼ĨZ֩ϧͨỈ Ö[ÝúăgƯ€"Þ`ö4ی̼ĩZÍ¨Ò Ï§Ó§ÌŒÍ¨Í6ͼڌĨ)Ñ

ø ÷Þ

P1 P2

P3 P4 O1

O3

G1

G2

G3 O2

ÎZǔÒ3ý)ÐLʆˆ̀Õ1ʼ۔֔թϧÎZÇ÷·å*è XêWשçUƆö ½שåâèíXêW;ø;ˆÉx̗ͨÎZϧϾɚـΠˆÍ˼ڌÍÔZ̗ӧՔչ̀Ì4ÓØÇGÎ̼É[ɈÈGÞ

1 PICK_GRIPPER_ball3_roomA 1 PICK_GRIPPER_ball3_roomA (left)

1 PICK_GRIPPER_ball1_roomA 1 PICK_GRIPPER_ball1_roomA (right)

1 PICK_GRIPPER_ball4_roomA 2 PICK_GRIPPER_ball3_roomA (free)

1 PICK_GRIPPER_ball2_roomA 2 PICK_GRIPPER_ball1_roomA (free)

2 MOVE_roomA_roomB 3 PICK_GRIPPER_ball4_roomA (left)

3 DROP_GRIPPER_ball4_roomB 3 PICK_GRIPPER_ball2_roomA (right)

3 DROP_GRIPPER_ball2_roomB 4 MOVE_roomA_roomB

3 DROP_GRIPPER_ball3_roomB 5 DROP_GRIPPER_ball4_roomB (left)

3 DROP_GRIPPER_ball1_roomB 5 DROP_GRIPPER_ball2_roomA (right) Level Abstract Plan Level Scheduled Plan with Free/Unfree(Realloc) Actions

C O N V E N T I O N A L

P L A N N E R

DECLARAT-IVE

SAT (BLACKBOX)

¿g‰OV%TpI5UIZm^LdKMŠP$I5]OI5G€I5KW£UV$O"Rb]U"P$VcU"V\I5O"RbKLdKMkHS—]G!m^LdTG&V–P0oRbL&P%V$O3I5UV˜I\iZI5L&G&I5xG&V

_oL&P0o.I5UV_O"]S3S3I5U"L& $V%WQL&K±…L&M5]UV Ç Ñ)ebš TUR5mR5m‰“T<V_L&STGdV$SV$K„m0IZmL&RbKtR5Ylæçè)éZêpébèZë

1 3 5 7 9

# robots

0.1 1.0 10.0 100.0 1000.0

Performance of Graphplan v/s Planning+Scheduling

in the Shuffle problem with varying number of blocks and robots

SHUF4-GP-TOT SHUF6-GP-TOT SHUF8-GP-TOT SHUF10-GP-TOT SHUF4-PS-TOT SHUF6-PS-TOT SHUF8-PS-TOT SHUF10-PS-TOT

Performance of Graphplan v/s Planning+Scheduling

in 8-Block Inversion Problem in Blocks World

GP-G GP-S GP-TOT PS-G PS-S PS-TOT

1 3 5 7 9

# robots

0.1 1.0 10.0 100.0 1000.0

Performance of Graphplan v/s Planning+Scheduling

in the 10-blocks huge fact and 9-block bw-large-a problems

HUGEFACT-GP-TOT HUGEFACT-PS-TOT BWLARGEA-GP-TOT BWLARGEA-PS-TOT

;7ÙÖ6È

1 3 5 7 9

# robots

0.1 0.8 8.0 80.0 800.0

Performance of Graphplan v/s Planning+Scheduling

in the Shuffle Problems in Blocks World

SHUF-4-GP SHUF-6-GP SHUF-8-GP SHUF-10-GP SHUF-4-PS SHUF-6-PS SHUF-8-PS SHUF-10-PS

S.IZL&KsO"L& $V%OR5Y'iZI5UL&I5xGdV$OžLdOžibV%U"‰c]OVY]GHL&KsTU^I5Pm^L&P%V5el±½R5UV‚I5STG&VZk<moV ðSÍa§“¬À¨

TUR5xG&V%S _Ldm^o`¸cUR5x<R5mOœ|I5OtO"RbGdi5V$WLdK`¸sO"V$P–LdKŠP%GJI5O"O—±…ŸzyƒYRbG&GdR_L&KMnm^oV–I5xERibV

RbUWV$U$kpx]m_m0I5~ZV$O,ѳhS3LdK“])m^V$O‡_oV$KcP%GJI5O"O¤Ÿ*#†‡•€Ó(Ƌ#Ô|I5O_OTEV$P$L!zpV$WŽ]TˆYURbK„m\e

ŸKÀ±…L&Mb]UVsÁ5ÁkggVvOV$VsmoVuTEV$U"YR5US3I5KP$VcR5Yhm^oVuKV SV%moR)WXRbKìm^oV‘³º7xGdR)P0~

1 3 5 7 9

# robots

0.1 1.0 10.0 100.0 1000.0 10000.0 100000.0

in the 10-block huge fact and 9-block bw-large-a problems

HUGEFACT-GP-TOT BWLARGEA-GP-TOT

G&RbM.OP\IZG&V5el±pRbU “¬

 )%ø“«úÏk<TUR5xG&V%S3O€_Ldm^ouU"RbxER5m^O—³œm^R˜Ì–I5U"V—L&Kv†|G&I5OO[Ÿ*#ž†‡•,Ó(Ưí#

_oL&GdVŽYRbU   ¥!ø5÷  ¨ ø5k'TU"RbxGdV$SO._LdmojURbxER5mOs³sm^R¸IZUVsLdK †|GJI5O"OnŸ*#†‡•€Ó(Ư‹#œe

2 3 4 5 6 7 8 2

7 1

10 100 1000

2 4 6 8 10

ĩzîòï_ðzóí2ô õÜóö_÷Û÷ ñÝøôí_ùµ÷Ýú í ðzûüñÝîÛ÷ ýzïuøýüñÝí2îÛö_õăñÝøï_ý õÜóö_÷Û÷ ñÝøôíüùµ÷Ûú âPđ
êPôMơ Ö2×ăØ2ÙÝÚÜÙÝØÛÞAßÝưÜì ìz÷Ûí2ị ó ôøñ

ỏ”ỉÒÉZơờỹớễÙẻ ẹÒốớẻxễ Iê Iạ ẹÒốớẻễ Iê Iạ ẹÒốừẻễ Iê I[ạ

à à ỉ—ũờđ”ÊỦ`ố¾ẹÒÚ ồ”ễ€ă{ịZũ”ỹớẹÉZỏ”ẹÒố¾ÉˆđßẺỚêỨỹ¾ẹÒêỨỹ¾ẹỉ—ũ”đờÊỦ`ố¾ẹÒÚ ẹÒڌễdêTS<U Ư ĩDV ạ.à çÓê`ö4ìậÒè‡ả

ả.èê-'WÒXFè]à àÉZỏ”ẹÒố¾Éˆđ”ÊÙẺèỨÉVạ ẻễÙịZđ”Êç]4‘Œ›@‡†ž`_v‡Z…ẵỡ Ẻ

IQK c]M]dD

LzJdR

NR9S”Ẻ

é¡îÛï_ðzóê2ô ñÝøôê íT+ í íè ñÛøôê íT+ í íè ñÛøôê íT+ í íè

÷E5ì¡ñòñÝóê 4_õ*)F÷) &)$ 4 &  4 4 )%% $! 4   &'4 ) 4 &

ìgΠ֩ϾÍhȒà

çg̼̀æ{ÉZ̗ÈxÎZǔىÍÜÉZæ à àÎZǔÒGê­Ï¾Ë¼ê­Ï¾ËÓ§ÇË¼ÚŒÍ ï Új۔ˆ˼ϾÍ.Ք̼ÉZ֩ϧͨÈxʀÞ

ð !$( !_ézîòï_ð& _î   &   ) &   4 &'4 

ð !$( !_ézîòï_ð& )Fî   &    & 4  4 &'4 

ð !$( !_ézîòï_ð& &'_î   &  & & 4  4 &'4 

ð !$( !_ézîòï_ð& $_î  ! &  0)) & 4  ! &'4 

ð !$( !_ézîòï_ð& _î  &' &  )$% 
 & 4   &'4 

ð !$( !_ézîòï_ð& 4_î  &' &  $4 %4 & 4   &'4 

ð !$( !_ézîòï_ð& )$_î  &'! &  
0$4  & & 4   &'4 

ð !$( !_ézîòï_ð& !_î  &'% &  äMï
,     &'4 

ð !$( !_ézîòï_ð& &'_î   &       &'4 

ìgΠ֩ϾÍhȒý

å`ͨʼ۔ϾË;Ó§Çx˼ڌͭթΠ̗ـç@ϧÎZÇ6֩ϾɚټäšÊUÑ;ÉZ̗ϧÒ6ҌɈÈxÎZÓ§Ç6Ñ`Ӿ˼Ú6È(۔ϧ˼ӾթϾÍ̼ÉZÖ¹ÉZ˼Ê4ÎZǔÒϧӧÈxӧˆͨÒ֩Ͼɚټä

չɈʼӾ˼ӧɈǔÊ;ɈÇ˼ڌÍ.˼Π֩ϧÍZÞ

šžG!m^šžG!m_I5GdORQOGdR7_O¤_L!m^oˆS3R5UV[UV%ORb]UP$V%O$e

‘ẻP’êỨỹớỹ¾ễÙđắÈ`UŒẺ¾ÈklFễÙđ”ồ”ỹ¾ễ€ỉ0ÈêU”Ẻ¾È]ịZđờồ<ìgị ẹ†ễZÈê(ẺƠ{ễÙồắẺĨụẵẺnm•ÒwV– yỎŠ[ẳ ẾyỎŠpoz}w ŠờŠ”yỎŠ[ẳV“ ạÉZỉÒԈịZđ

ỎịZũ”ă{ẻxịZđờđGçUũŒơờỹằẺ¾È ĩ ịZđ5ạ5ị ẹ†ễ€ÉŒÈắö4ê(Ẻ'ẻQHQM1šẺ

Újẹ†ẹ†ỏ cLỹHỹ)ỦỦỦưẺ‚ứ€Ê€Ẻẹ†ÉZỉÒɈđjẹ†ÉŒẺễÙồ”ũbỹịZố¾ỏờÊ"àM1H1H1aỹ ĩ ễÙỹ¾ăằö;ɈđVẹÒịZốừđŒễÙồ”êFậpç ĩ ƯỹàM1H1H1šẺỏờʀẺỲàM1H1H1šẺ

‘Ỉ/’ă”Ĩ¨â¹×VơrÞ¾× ÎZ̔Ò8-íĨˆÌŒÔŒ×MS"Þ3¬a‹Œ•@˜•€›L•—¡¨›œy·‡ Š6w ŠQ–­u}—’z¾‡ yœ›¢wZ›œyœ‡ZŠ‡Ož4 ”« vv•€›œ…’«y{Š˜o4z¾w Š”Š”y{Šj£

˜"‡ vxw y{ŠŒ‰“4ìÙ¼ĨšÙ ތƯ‰U[ö4í`Ư™ÐvQHQ[Þ²ỈQHQHQ[Þ

‘ỈQ’™lFĨ þßỈxÎZÌß× UŒÞ¾× ÎZÌ©Ò ôFÍ€Ö¹Í¨Ï·× ÿ­Þ ¬a‹š•šu›u oz}w Š©Š”y½Š[£  ”«’›L•‰vx|œủw‰›oz}w Š

s •€ŠQ•‰…†wZ›œy·‡ Š ¬a‹V…†‡ ‘ˆ£ ‹ž†"•‰‘š…¤y½’›œy·¡ú Q•¼w …†¡‰‹š“ ï ۔֩ỈxÓ¾Ị†Ị†Í¨Ò Ị†Ĩ UˆíFƯ¡ơÜÞ(í­Ï§Í†Ĩ Ñ;̀֩ͼӾỊ†Í Î Ị

Þ ỉ]Ì Ị¼ÚŒÍƒơ`ͨϧΠỊ¼Ó¾ĨˆÌ”ÍǹֹỊLÑ;Í€Í¨Ì ƯOÌVỊ†Í¨ÏØϧӾÔZͨÌjỊ ÿ4ÎZÙ¼ô[Ị†Ì—ÎZÙ¼ôšÓ§ÌŒÔ

ÎZ̔Ò$ă”ÎZÓØϧی̼͉Ð}S­Ì—Ó¾ĩZͨÌWâšÕ©Ï§ÎZÌ©Î Ị¼Ó¾ĨˆÌšÐ¢Ö©ÎZ͆ͨÒ-íͨΠ̤̔ӧ̌ÔtÓ§Ì ö;ĨˆÌ©Í†Ị†Ì—ÎZÓ§ÌjỊ ï Î Ị¼Ó§Í†ưœÎZىỊ¼Ó¾ĨˆÌ ÎZÌ©Ò

ì@ϧÎZÌ”Ì”Ó§Ì”ÔŒÞ ¥ …’›œy”@¡€y·w z q Š›L•‰z{z¸y¾£j•€Š¡¤•€Þ'ỈQHQHQ[Þ

ăƯ vFŒ•Ẩ›œy{›œy·Ư Àjm•’’‘Ãz›{€“rở4ậỗ½ậ†ếaނỉ€ấ€Þ݈ẽZé¾ẻZÞẻẨề”íbé)ếíŒệ"éẩxỉ€ềŒẻ€èôẩẫZậ†ậKéẽZể¾ếẪấỊỉ‰ẫˆẩxếŒÐđèỊẻẨấỊí”éưậỊấ€Þ‚ĩVậỊẩé·Þ

Performance of Graphplan v/s Planning+ Scheduling

in 8-Block Inversion Problem in Blocks World

GP-G GP-S GP-TOT PS-G... class="page_container" data-page="46">

1 9

# robots

0.1 0.8 8.0 80.0 800.0

Performance of Graphplan v/s Planning+ Scheduling

in. .. class="page_container" data-page="43">

1 9

# robots

0.1 1.0 10.0 100.0 1000.0

Performance of Graphplan v/s Planning+ Scheduling