Laporan Akhir H ibah P e nelit ian T im P as c as ar jana – H P T P (H ibah P as c a) P ENGEMB ANGAN P E MODE L AN B ERBAS ISK AN A GEN YANG MELIBATKAN EMOSI D AN PEMBEL AJ AR AN Periset Utama Dr. Ir. Utomo Sarjono Putro, M.Eng Riset ini dibiayai oleh ITB berdasarkan Surat Perjanjian Pelaksanaan Riset No: 0407/K01.03/Kontr-WRRIM/PL2.1.5/IV/2008 Lembaga Penelitian dan Pengabdian Pada Masyarakat INSTITUT TEKNOLOGI BANDUNG 2008 Ekonomi, Ilmu Sosial dan Politik
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Citarum was a clean river where local people enjoyed fishing and recreation, however now
its condition has already changed totally. Currently, the river can not provide its social
services, such as clean water, electricity, fishing, tourisms, transportation, and public
recreation. In rainy season, the color of river is brown because each drop contains mud from
bald lands erosion along the river. In dry season, the color of river is black and full of
household wastes. In spite of many seminars were held, classical problems always arise.
There are floods when rainy season and drought when dry season. There are some factorswhich cause the problem, i.e.; illegal lodging and the population exploding in upper stream,
pollution from industries in down stream, etc. This paper tries to see the Citarum problem
from a different point of view, starting from the belief that if stakeholders (or agents) in the
Citarum problem collaborate to solve the problem, and then Citarum river will be better and
better.
In previous research we had adopted PAD emotional model and drama theory in order to
develop an agent based negotiation simulation. Our previous simulation was successfully
showed that agent’s positive emotion will affect their ability to eliminate confrontation
dilemmas. But unfortunately this simulation was unable to analyze collaboration dilemmas
among agents. The purpose of this research is to develop more advance simulation that is
also able to analyze collaboration dilemmas (especially trust dilemma) among agents.
Keywords: Agent based Simulation, Negotiation, Dilemma, Drama Theory, Emotion.
Namun, simulasi yang telah dikembangkan terdahulu tidak mampu menganalisa dilema-
dilema yang dapat muncul pada tahap kolaborasi.
Tujuan penelitian ini adalah mengembangkan model simulasi negosiasi yang
tidak hanya mampu menganalisa dilema-dilema konfrontasi tetapi juga, mampu
menganalisa dilema-dilema kolaborasi khususnya, trust dilemma. Dalam rangka
membangun model barui ini, diadopsi konsep tendensi afiliasi yang mencerminkan
kecenderungan suatu agen untuk bekerja sama dan berkolaborasi dengan agen-agen lain.
III. PERNYATAAN MASALAH, TUJUAN RISET, METODOLOGIDAN REFERENSI
III.1 PERNYATAAN DAN TUJUAN RISET
Sungai Citarum pada jaman dahulu mampu menyediakan jasa-jasa sosialnya
berupa air yang bersih, pembangkit listrik, perikanan, pariwisata, dan rekreasi umum
lainnya. Namun saat ini, sungai Citarum sudah tidak mampu menyediakan jasa-jasa itu
dengan baik, malahan merupakan sumber bencana bagi masyarakat sekitar berupa banjir
pada musim hujan dan juga berbagai penyakit karena kotornya air sungai pada musim
kemarau.
Banyak pihak yang terlibat dalam permasalahan tersebut, yaitu penduduk dan pemerintah daerah di daerah hulu, hilir, industri, dan lembaga swadaya masyarakat.
Beberapa usulan untuk mengembalikan fungsi sungai Citarum sudah ada, namun belum
ada suatu usulan yang komprehensif sehingga semua pihak dapat bekerja sama untuk
memecahkan masalah tersebut.
Permasalahan sungai Citarum ini sudah berlangsung sangat lama, dan sudah
berpuluh-puluh seminar dilaksanakan sehingga diasumsikan semua pihak sudah saling
berinteraksi lama dan sudah saling mengenal dan berbagi informasi. Oleh karena itu,
dalam analisis hambatan kerjasama dengan drama theory ini, akan mulai dari tahap akhir
dari build-up, di mana semua pihak sudah mempunyai common reference frame terhadap
permasalahan sungai Citarum. Pertama-tama, common reference frame awal dibuat
dengan berdasarkan atas informasi yang didapat dari surat kabar dan beberapa bahan dari
seminar mengenai DAS Citarum. Kemudian, analisis akan memasuki tahap climax
Seperti yang dapat dilihat pada Gambar 1, proses negosiasi menurut drama theory
akan terdiri dari lima tahap, yaitu:
1. Scene Setting
Dalam tahap ini, pihak-pihak yang berinteraksi belum mempunyai persepsi yang
sama atas situasi yang dihadapi (atau, belum mempunyai common reference),
dimana masing-masing pihak berusaha mempengaruhi pihak lain dengan frame
nya, atau pun dia belajar tentang frame yang dilihat pihak lain. Setiap pihak bisa
melihat situasi dengan frame yang berbeda, sehingga ada beberapa alternatif
frame yang tersedia pada tahap ini. Tahap ini akan menghasilkan beberapa frame
yang bisa berbeda satu dengan yang lainnya.
2. Build-up
Tahap ini mulai dengan kondisi di mana tiap pihak mempunyai frame yang
berbeda, dan masing-masing pihak berkomunikasi dengan pihak lain, sehingga
ada kemungkinan kesalahan persepsi terhadap frame pihak lainnya. Situasi ini
disebut dengan kondisi hyperframe. Namun dengan adanya interaksi dan
komunikasi di antara mereka, maka lambat laun setiap pihak akan mempunyai
frame yang sama terhadap situasi konflik yang mereka hadapi, sehingga mereka
mempunyai common reference frame. Dengan frame yang sama, maka tiap pihak
mulai bisa menawarkan posisinya masing-masing secara terbuka. Sering dalamtahap ini, posisi yang ditawarkan berbeda satu pihak dengan pihak lain, atau pun
kalau mereka menghasilkan posisi bersama, maka ada kemungkinan masing-
masing tidak percaya dengan komitmen pihak lain terhadap posisi bersama
tersebut. Kalau mereka menghasilkan posisi bersama dalam tahap ini, dan mereka
juga tidak mempunyai masalah dengan kepercayaan dengan pihak lain, maka
proses konflik bisa langsung meloncat ke tahap 4, yaitu resolusi, dimana mereka
akan siap menghasilkan suatu penyelesaian, happy ending or tragedy.
3. Climax
Tahap ini bermula ketika ada pihak yang tidak puas dengan frame bersama dan
berusaha untuk mengubahnya. Dalam hal ini, frame yang sama tersebut tidak
menghasilkan posisi bersama, atau pun kalau menghasilkan posisi bersama,
masing-masing pihak masih ada ketidakpercayaan dengan pihak lain. Dalam
seharusnya menghentikan pembuangan sampah ke sungai, USP seharusnya
menghentikan penebangan liar, dan DSC seharusnya tidak memberlakukan hukuman
keras pembuangan sampah/limbah ilegal ke sungai, tidak seharusnya merawat secara
tuntas, dan tidak seharusnya membagi pendapatannya ke USR.
Penjelasan posisi ancaman (Threat )
Posisi ancaman akan terealisasi bila masing-masing pihak menjalankan ancamannya.
Untuk kasus di atas, posisi ancamannya adalah USR tidak akan menghentikan
deforestation, Green akan memprotes, TI tidak akan menghentikan pembuangan limbah
yang belum diolah, DSP tidak akan menghentikan pembuangan sampah rumah tangga,
USP tidak akan menghentikan penebangan liar, dan DSC akan memberlakukan hukuman
yang keras terhadap pembuangan sampah/limbah ilegal ke sungai, tidak akan merawat
dengan tuntas, dan tidak akan membagi pendapatannya ke USR.
2. Model Simulasi Teori Drama
Untuk dapat membawakan suatu permasalahan teori drama ke dalam simulasi
dibutuhkan definisi-definisi yang berhubungan dengan pembentukan kerangka bersama
permasalahan. Dalam model simulasi ini diasumsikan bahwa tahap pertama teori dramayaitu scene setting sudah dilalui. Setiap agen sudah memberikan opsi yang ditawarkannya
sehingga dapat dipilih oleh agen lain.
Definisi 1:
kio adalah opsi agen-i yang dapat ditawarkannya kepada agen lain dengan k adalah
jumlah opsi yang dimiliki agen-i.
Untuk setiap opsi-k yang dimiliki oleh agen-i, agen-i dapat mempunyai pilihan menerima
ki → adalah pilihan yang dilakukan oleh agen-i terhadap opsi-k
yang ditawarkannya.
Penerimaan atau penolakan terhadap opsi-k yang dipilih oleh agen-i didasarkan padasuatu nilai yang diberikan oleh agen-i pada waktu-t untuk pilihan menerima atau menolak
opsi-k tersebut.
Definisi 3:
ℜ→)(: ki
t
ki
t
ki ocVo adalah nilai yang diberikan terhadap pilihan opsi –k oleh agen-i pada
waktu-t yaitu
⎪⎩⎪⎨⎧
−=
ditolak oc jikaa xditerimaoc jikaaocVo
ki
t
ki
ki
t
kiki
t ki
t ki
)(;)(;))((
x adalah nilai dasar
Untuk setiap opsi-k yang ditawarkan oleh agen-i, agen-j memberikan pilihan
terhadap opsi tersebut, diterima, ditolak atau tidak memilih pada waktu-t.
Definisi 4:
},,{: memilihtidak ditolak diterimaoc kij
t
kij → adalah pilihan yang dilakukan oleh agen-i
terhadap opsi-k yang dimiliki oleh agen-j.
Penerimaan atau penolakan yang dilakukan oleh agen-i terhadap opsi-k yang ditawarkan
agen-j didasarkan pada suatu nilai persepsi yang diberikan oleh agen-i pada waktu-t.
Definisi 5:
ℜ→)(: kij
t
kij
t
kij ocVpo adalah nilai persepsi opsi yang diberikan agen-i terhadap pilihan
opsi –k yang ditawarkan oleh agen-j pada waktu-t yaitu
Untuk setiap agen-i akan menentukan dilema terhadap setiap agen-j yang didasarkan
pada nilai persepsi posisi dan nilai persepsi ancaman sedemikian sehingga:
1. Jika nilai persepsi pilihan agen-j )( t
j
t
i pVpp lebih besar atau sama dengan nilai
persepsi ancaman )( t t
i t Vpt maka agen-i lebih menyukai posisi agen-j sehingga
dilema yang terbentuk adalah dilema penolakan (rejection dilemma).
2. Jika nilai persepsi pilihan agen-j )( t
j
t
i pVpp lebih kecil atau sama dengan nilai
persepsi ancaman )( t t
i t Vpt maka agen-i lebih menyukai posisi ancaman sehingga
dilema yang terbentuk adalah dilema persuasi ( persuasion dilemma).
Model simulasi teori drama ini akan membentuk suatu kerangka bersama dan dilema
yang muncul dari kerangka bersama tersebut. Dalam model simulasi ini hanya 2 dilema
yang akan dihilangkan yaitu dilema penolakan (rejection dilemma) dan dilema persuasi
(persuassion dilemma).
3. Model Negosiasi
Dalam model ini, proses negosiasi melibatkan unsur strategi dan keadaan emosidari agen dalam bernegosiasi. Unsur strategi akan menentukan seberapa besar keahlian
seorang agen dalam melakukan penawaran kepada agen lain. Keadaan emosi akan
menentukan seberapa besar pengaruh prilaku (cara berbicara, ekspresi, dan keadaan
mental) seorang agen dalam melakukan negosiasi. Hal ini memungkinkan seorang agen
dengan keahlian negosiasi yang rendah tetapi memiliki pengaruh prilaku yang baik dapat
mempengaruhi pihak lain dan pengaruh ini juga bergantung dari keadaan emosi pihak
lain.
Setiap agen yang memiliki dilema terhadap agen lain akan melakukan negosiasi
yang didasarkan pada strategi negosiasi yang digunakannya dan keadaan emosi yang
dimilikinya. Tujuan dari proses negosiasi ini adalah untuk mengajak agen lain merubah
pilihannya terhadap suatu atau beberapa opsi dimana pada opsi-opsi ini mereka memiliki
pilihan yang berlawanan (konflik). Seorang agen akan merubah pilihannya terhadap suatu
Agen-1 memiliki ketidakcocokan dengan agen-2 pada opsi O12, maka nilai
persepsi agen-1 terhadap pilihan Menerima (pilihan agen-1) akan berkurang 5 (87-5=82)
dan terhadap pilihan Menolak (tidak dipilih agen-1) akan bertambah 5 (13+5=18).
Misalkan keadaan emosi agen-1 adalah
}2.0,1,5.0{1 −= Es , maka fungsi keadaan emosi agen-1 adalah
2.12.0)11(5.0),,(1 −=−+×−=d a p r r r Se
Dengan adanya keadaan emosi, maka pengaruh negosiasi ini akan bernilai
1552.11111
−=+×−=+×= st st SeOv
Agen-j memiliki keadaan emosi },,{ d a p j r r r Es = yang akan mempengaruhi agen-
j dalam menerima tawaran dari agen-i. Dengan fungsi keadaan emosi agen-j
d a pd a p j r r r r r r Se −+= )1.(),,( , agen-j akan menilai penawaran agen-i dengan tingkat
penawaran yang dirasakan agen-j yaitu
ii j j OvOvSeOv +×=
Pengaruh nilai negosiasi ini akan mengurangi nilai persepsi pilihan agen-j terhadap nilai
persepsi pilihan yang dipilihnya dan menambah nilai persepsi agen-j terhadap nilai
persepsi pilihan yang tidak dipilihnya pada opsi dimana agen-i dan agen-j tidak cocok.
Proses negosiasi ini akan berlaku juga kepada agen-j yang akan bernegosiasi
dengan agen-i (kebalikannya). Dalam hal agen-i melakukan negosiasi terhadap agen-j,
agen-i berperan sebagai pihak penawar yang memiliki strategi dan agen-j berperan
sebagai pihak penerima tawaran. Proses negosiasi ini akan berlangsung sampai suatu
waktu-t tertentu dimana agen-i yang berperan sebagai pihak penawar akan berhasil dan
agen-j sebagai pihak penerima tawaran akan menerima tawaran agen-i. Hal ini akanterjadi jika nilai persepsi agen-j pada waktu-t terhadap opsi yang dipilihnya dari t=0
sampai dengan t-1 sudah lebih kecil dari nilai persepsi agen-j pada waktu-t terhadap opsi
yang tidak dipilhnya dari t=0 sampai dengan t-1. Hal ini akan merubah pilihan agen-j
terhadap opsi dimana agen-i dan agen-j memiliki ketidakcocokan.
Abstract. The purpose of the present research is to identify, analyze and simulate dynamics of interaction and conflicts among agents using drama theory, and to apply it in Citarum river basinproblem. To accomplish the purpose, we first model the process in terms of drama theory that is
combined with emotional state model (PAD). One of the reasons why we adopt drama theory is that itprimarily focuses on dilemma or paradox arising from rational goal seeking behavior. It also providesus with rigorous analytical and computational tools of conflict analysis. Then, we propose a simulationmodel to describe both of dilemma of conflict, i.e., persuasion and rejection dilemma among the agents,
and the dilemma of collaboration (trust dilemma) among the agents. Finally, we conduct agent-basedsimulation by using SOARS (Spot Oriented Agent Role Simulator ) to obtain some fruitful suggestionsfor encouraging their collaboration.
Keywords: Agent based Simulation, Negotiation, Dilemma, Drama Theory, Emotion
1 Introduction
Citarum River basin is a region in Java Island, Indonesia, with 6,080 km2
area involvingthe three provinces, i.e., West Java, Banten, and Jakarta. The annual precipitation depth is
3,000 mm/year in the mountain and 2,500 mm/year in lowland. Relative humidity is 80%and daily temperature is 25○C in the lowland and 18○C in the mountain. Citarum River is
connected with 4 rivers to the west and 4 rivers to the east in West Java. In the past,
Citarum river was clean, but now the condition had changed totally [17].
Fig. 1. Citarum River Basin Region
There are some factors which cause the problem, i.e.; illegal lodging and the population
explosion in upper stream, and household waste in down stream. Agent in the Citarum problem has different interests and positions. There are several agents who participate in
Citarum river basin, i.e. local people in downstream, local people in upstream, textile
industries, environmentalist (green), regencies in upper stream and cities in down stream[19]
In negotiation process, agents may change their position and interests; accordingly, the
situation is dynamic. There are some impediments (or dilemma) to achieve common
position and trustworthy (i.e., collaboration). Based on the previous research [16], theeffect of positive emotional state of agent is important to make negotiation. The results
from simulation show that number of dilemma among agent could be reduced
significantly or even have no dilemma.
In this research, we add a conceptual model for affiliation that is affiliate tendency. It wasdefined in terms of generalized positive social expectations and associated positive
behaviors in social interactions with others.
Although dilemma of confrontation has been eliminated, dilemma of collaboration oftenstill remains. Affiliate tendency is tendency to cooperate with other agent. It causes a new
problem among the agent that is trust dilemma. How does each agent eliminate the trust
dilemma? This paper will propose a simulation model of negotiation process involving
dilemma of conflict and trust among the agents. Then we conduct agent-based simulation by using SOARS (Spot Oriented Agent Role Simulator ) to obtain some fruitful
suggestions for encouraging their collaboration.
2 Drama Theory in Citarum River Basin Problem
Drama Theory is a metaphor of dynamic confrontation process. This paper adopts the
metaphor, and starts with the build up stage of interaction among agents in Citarumriverbasin problem. Common reference frame resulted from the stage is described by Fig.
3. The common reference frame consists of agents/participants, their options, positions
Fig. 3. Common references frame in Citarum river basin problem
3 Agent Based Modeling in Drama Theory
The purpose of this paper is to identify, analyze and simulate dynamics of interaction and
conflicts among the stakeholders (or agents) in the Citarum river negotiation process.
They claim their strategies and interests as well as express emotion. To accomplish the purpose, we first model the process in terms of drama theory that is combined with
emotional state model (PAD). One of the reasons why we adopt drama theory is that it primarily focuses on dilemma and also provides us with rigorous analytical andcomputational tools of conflict analysis. Then we conduct agent-based simulation by
using SOARS. In this paper, we only discuss emotion model and temperament model
(affiliative tendency). The detail of agent based modeling in drama theory could be find
in the previous research [16].
3.1 Emotion model and Affiliative Tendency
Emotion model that will be used in this paper is the development from emotional
negotiation model PAD [9]. Emotional state model (PAD) involves three dimensions, i.e.,
Pleasure (r p), Arousal (r a) and Dominance (r d ). During negotiation, a more pleasantagent tends to compromise with others. Each agent has the emotional state [9], i.e.:
},,{ d a pi r r r Es = ; )1,1(,, −∈d a p r r r (1)
Based on the previous paper [11], angry is coded by {-.51, .59, .25}, bored is {-.65, -.62,-.33}, curious is {.22, .62, -.01}, dignified is {.55, .22, .61}, elated is {.50, .42, .23},
hungry is {-.44, .14, -.21}, inhibited is {-.54, -.04, -.41}, loved is {.87, .54, -.18}, puzzled
is {-.41, .48, -.33}, sleepy is {.20, -.70, -.44}, unconcerned is {-.13, -.41, .08}, and
violent is {-.50, .62, .38. Each agent i has the function of emotional state [9], which is:
d a pd a pi r r r r r r Se −+= )1.(),,( (2)
For example, if an agent has emotional state defined as {−0.51, 0.59, 0.25}, then function
of emotional state is 0609.125.0)59.01.(51.0)r ,r ,r (Se d a pi −=−+−= Affiliate tendency was defined in terms of generalized positive social expectations and
associated positive behaviors in social interactions with others. An individual’s emotional
states are inferred from averages of his or her emotional states across representativesamples of everyday situation. Thus, the previous paper [12] proposed that emotional
traits could also be described in terms of the pleasure-displeasure, arousal-nonarousal,
and dominance-submissiveness dimensions. Then, affiliate tendency scales were definedas follows:
ijijijij rd rarpn Affiliatio 3.024.046.0 ++= (3)
Within the present theoretical perspective, then, it is important to conceptualize andmeasure affiliate tendency as pure generalized interpersonal positive ness without either
an inclination to want to dominate and control others or to be dominated and controlled
by others.
3.2 Modeling to Eliminate Trust Dilemma
To eliminate trust dilemma in collaboration stage, we use affiliate tendency value. First,we calculate probability of cheating and punishing another agent who cheated.
minmax
max)(Pr V V
aff V iob i
−
−=
(4)
maxV is maximum value of affiliate tendency (1.00); minV is minimum value of affiliate tendency (-1.00)
and iaff is affiliate tendency for agent i. The higher probability is the higher probability of the agent to
punish another agent who cheated. In contrast, the lower probability is the higher probability of the agent to
cheat.
Based on norm game [1], agent i will attempt to cheat the commitment if rand iob <)(Pr ,
where rand is random value which is generated from uniform distribution in range [0,1].
For each agent j ( i j ≠ ), he/she will attempt to punish agent i if rand job >)(Pr . If agent
i cheat, then agent i’s payoff will increase by 1 %, while the payoff of other agents willdecrease by 1%. If agent i cheat and agent j punish, then payoff of agent i will decrease
by 10%.
4 Simulation Using SOARS and Result
In order to simulate this problem, we use SOARS to describe the initial frame for
Citarum river basin problem as seen in figure 3. There are so many dilemmas in the
common frame. Based on the previous research [16], we can eliminate the dilemma of
conflict. In this current paper, we assume that bargaining strategy of agent was same, that
is sst i = ts∈ . We make four experiments to look measure level of emotional state, so the
agent could negotiate in order to reduce dilemma and to eliminate trust dilemma.
4.1 First Scenario
In this scenario, whole of agents have positive emotion. USR is having a strong desire to
know about something, G is having strong feeling of deep affection for something, TI isexcitement, USP is calm, DSP is quiet and DSC is calm. In this scenario, we use the
parameters as described in Table 1, and then the probability of affiliate tendency could be
seen in figure 4.
Table 1. Parameters in Scenario 1
USR G TI DSP USP DSC
pr 0.22 0.87 0.5 0.2 0.55 0.55
ar 0.62 0.54 0.42 -0.7 0.22 0.22
d r -0.01 -0.18 0.23 -0.44 0.61 0.61
Sei 0.3364 1.5198 0.48 0.5 0.0610 0.0610
Fig. 4. Probability of Affiliate tendency in Scenario 1
From figure 4, we can see that the probability of affiliative tendency for each agent is
close to 1, it means that no agent will cheat the commitment, because affiliate tendency is
positive and the value is large. Each agent will avoid cheating because the other agent can
punish him/her.
4.2 Second Scenario
In this scenario, number of agent who has positive emotion is more than number of agent
who has negative emotion, and level of emotional state for agent with negative emotion is
low. USR, USP, DSP, DSC and TI have positive emotion, meanwhile G has negativeemotion. USR and DSC are excitement, USP and DSP have strong feeling of deep
affection for something and TI is calm, G is not worried. In this scenario, we use the
parameters as described in Table 2, and then the probability of affiliate tendency could be
seen in figure 5.
Table 2. Parameters in Scenario 2
USR G TI DSP USP DSC
pr 0.5 -0.13 0.55 0.87 0.87 0.5
ar 0.42 -0.41 0.22 0.54 0.54 0.42
d r 0.23 0.08 0.61 -0.18 -0.18 0.23
Sei 0.48 -0.1567 0.061 1.5198 1.5198 0.48
Fig. 5. Probability of Affiliate tendency in Scenario 2
From figure 5, we can see that the probability of affiliative tendency for each agent is
close to 1, it means that no agent will cheat the commitment, because the affiliatetendency is positive and the value is large. Each agent will avoid cheating because the
other agent can punish him/her.
4.3 Third Scenario
In this scenario, number of agent who has positive emotion is more than number of agentwho has negative emotion, and level of emotional state for agent with negative emotion is
moderate. USR, USP, DSP, DSC and TI have positive emotion, meanwhile G has
negative emotion. USR and DSC are excitement, USP and DSP have strong feeling of
deep affection for something and TI is calm, and G has a strong desire for something. Inthis scenario, we use the parameters as described in Table 3, and then the probability of
Fig. 6. Probability of Affiliate tendency in Scenario 3
From figure 6, we can see that the probability of affiliative tendency for each agent isclose to 1, it means that all agents will keep their commitment, because affiliate tendency
is positive and the value is large. Each agent will avoid cheating because the other agent
can punish him/her.
4.4 Fourth Scenario
In this scenario, number of agents who have positive emotion is more than number of agents who have negative emotion, but level of emotional state for negative emotion is
high. USR, USP, DSP, DSC and TI have positive emotion, meanwhile G has negative
emotion. USR and DSC are excitement, USP and DSP have strong feeling of deepaffection for something, TI is calm, and G is angry. In this scenario, we use the
parameters as described in Table 4, and then the probability of affiliate tendency could be
seen in figure 7.
Table 4. Parameter in Scenario 4
USR G TI DSP USP DSC
pr 0.5 -0.5 0.55 0.87 0.87 0.5
ar 0.42 0.59 0.22 0.54 0.54 0.42
d r 0.23 0.25 0.61 -0.18 -0.18 0.23
Sei 0.48 -1.045 0.061 1.5198 1.5198 0.48
Fig. 7. Probability of Affiliate tendency in Scenario 4
From Figure 7, we can see that the probability of affiliative tendency for each agent is
close to 1, it means that no agent will cheat the commitment, because affiliate tendency is
positive and the value is large. Each agent will avoid cheating because the other agent can
punish him/her.
4 Conclusion
From the results of simulation, we showed how the emotional states and affiliate
tendency of agents significantly affect the negotiations process. In the collaborationstage, to maintain the commitment, each agent must be willing to punish the other agent
who attempt to cheat. The effect of positive emotional state of agent and the affiliate
tendency is important to maintain collaboration.The results suggests that each agent should have a positive emotion and positive affiliate
tendencies in order to achieve collaboration in the Citarum riverbasin problem.
References
1. Axelrod, R. (1997), The Complexity of Cooperation. Agent-Based Models of Competition and Collaboration,Princeton University Press, Princeton, New Jersey.
2. Bryant, J. (2003), The Six dilemmas of Collaboration, John Wiley.
3. Barteneva, Daria., Lau,Nuno., Reis, Paulo Luis., (2006), “Implementation of Emotional Behaviors in MultiAgent System Using Fuzzy Logic and Temperamental Decision Mechanism”, Proceeding Fourth EuropeanWorkshop on Multi Agent Systems.
4. Bradley, Margaret M., Codispoti , Maurizio., Sabatinelli Dean., dan Lang, Peter J. (2001), “Emotion and
Motivation II: Sex Differences in Picture Processing” , Emotion, Vol.1, No.3 pp:300-319.5. Howard, N. (1996), “Negotiation as Drama: How “Games” Become Dramatic”, International Negotiation, 1, pp.
125-152.6. Howard, N., Bennet, P., Bryant, J. and Bradley, M. (1993), “Manifesto for a Theory of Drama and Irrational
Choice”, Systems Practice, 6(4), 429-434.7. Howard, N. (1994a), Drama Theory and its Relation to Game Theory: Part One”, Group Decision and
Negotiation, 3, 187-206.8. Howard, N. (1994b),”Drama Theory and its Relation to Game Theory: Part Two”, Group Decision and
Negotiation, 3, 207-235.9. Jiang, Hong, Vidal, Jose M and Huhns, Michael N, (2004), Incorporating Emotions into Automated Negotiation.
University of South Carolina, Columbia.10. Maulana, F. (2004), Menyelamatkan Hutan Tatar Sunda, Kompas Online 12 Mei.11. Mehrabian, Albert. (1976). “Questionnaire measures of affiliative tendency and sensitivity to rejection”.
Psychological Reports, Vol.38, pp:199-209.12. Mehrabian, Albert. (1997).”Analysis of Affiliation Related Traits in Term of PAD Temperament Model”. The
Journal of Psychology Vol.131, pp:101-117.13. Putro, Utomo Sarjono, (2000), Adaptive Learning of Hypergame Situations Using a Genetic Algorithm, IEEE
Transactions on Systems, Man, and Cybernetics, Vol. 30, No. 5.14. Putro, Utomo Sarjono, et al., (2005), Agent Based Modeling and Simulation of Knowledge Management,
Proceeding IFSR.15. Putro, Utomo Sarjono, et.al. (2005), Drama Theory sebagai Model dari Dinamika Konflik dalam PermasalahanDAS Citarum, Jurnal Manajemen Teknologi, Vol. 4, No. 2.
16. Putro, Utomo Sarjono, et.al. (2007), Role of Emotion in Negotiation Process: An Application of Drama Theory
in Citarum River Basin Problem, International Society of System Science.