An Electricity Market Game using Agent-based Gaming Techniq … · 2018. 10. 24. · An Electricity Market Game using Agent-based Gaming Techniq ue for Understanding Energy Transition

An Electricity Market Game using Agent-based Gaming Technique forUnderstanding Energy Transition

Setsuya Kurahashi1 and Wander Jager2

1Graduate School of Business Sciences, University of Tsukuba, 3-29-1 Otsuka, Bunkyo, Tokyo, Japan2University College Groningen, Hoendiepskade 23/24, 9718 BG Groningen, The Netherlands

Keywords: Electricity Market, Two-sided Market, Agent-based Gaming.

Abstract: The Electricity Market in Japan has been an oligopolistic market since the previous century, but it will bea liberalised competitive market soon due to a policy change. It is supposed to provide wholesale powermarkets. Therefore, it has high possibilities to become two-sided markets with strong wholesalers. The goalof this study is to clarify decisive factors for making decision of energy selection based on human competitiveand collaboration behaviour to be helpful for an incentive design of energy markets. For the purpose, twohypotheses were set in the experiment. First is that energy transition to renewable source is achieved byplayers while keeping their profit. Second is that aggregators have ability to control the energy market throughthe share of consumers’ power market as well as other two-sided markets. Our experiment confirmed thatthe energy orientation of electric power consumers could give a significant influence on power generationinvestment of electric power suppliers, and the risk of nuclear energy was underestimated. And the firsthypothesis was adopted and the second was rejected by the experiments through the agent-based gaming.

1 INTRODUCTION

The electricity crisis caused by the huge earthquakein Japan 2011, clarified that traditional electricity sys-tems on a one-sided energy market are inadequate formaintaining safe and stable electricity supply at lowcost. Given such an issue, the government of Japanhas clearly announced that it would realise liberationfor participation of power operators into small con-sumers such as general households in 2016. It wouldlaunch unbundling of power generation and distribu-tion during the period around 2018 to 2020. Thesepolicies might bring about advancement of innovationwith a wide variety of enterprises participating and in-creasing the use of renewable energy.

This attempt can encourage a wide variety of en-terprises into this market; however, it also entailssome risks such as instability of electricity marketsand market monopolies or oligopolies. These are dueto a two-sided energy-market on a de facto standardplatform as well as e-tailer and e-marketplaces.

The purpose of this research is to achieve an ef-ficient market while taking into consideration elec-tricity market liberalisation. Additionally, this re-search studies incentive mechanisms for a competitiveelectricity markets for enabling energy transformation

from fossil energy to renewable energy. In this re-search, social systems and infrastructures are referredto as the electricity market platform. Here, the fo-cus is placed on aggregators that bring electricity con-sumers together as a community. And it is also fo-cused on imbalance settlement which is implementedamong power distribution operators and power pro-ducers/retailers for the purpose of supply and demandadjustments for renewable energy.

Through this research, by applying the agent-based gaming method, our goal is to propose an in-centive design. It promotes innovation such as elec-tricity supply and demand adjustments, stable sup-ply, and dissemination of renewable energy throughfree decision-making by market participants includ-ing consumers and power operators. In a new lib-eralised energy market old and new energy compa-nies will base their actions and plans on the behaviourof their competitors as well as on the (expected) re-sponses of the consumer market. We propose usingan agent based simulation of a market of consumersas a laboratory setting to study the behaviour of hu-man decision-makers in an energy transition game.

314Kurahashi S. and Jager W.An Electricity Market Game using Agent-based Gaming Technique for Understanding Energy Transition.DOI: 10.5220/0006247703140321In Proceedings of the 9th International Conference on Agents and Artificial Intelligence (ICAART 2017), pages 314-321ISBN: 978-989-758-219-6Copyright c© 2017 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved

2 RESEARCH BACKGROUND

As clearly demonstrated by examples of the com-munications and Internet markets, the liberalisationof participation by many enterprises can create newmarkets while bringing about many benefits such asincrease in business opportunities, diversification ofservices, and lowering of fees. On the other hand,leaving everything to free market competition pre-vents products and services with higher transactioncost from being transacted. It results in causing mar-ket failure. Renewable energy is easily affected by thenatural environment, making the supply-and-demandbalance difficult to adjust, while power generationcost at the same time is expensive. This increasestransaction costs. Therefore, it is overly optimistic tobelieve that the price mechanisms within the marketcould for sure promote and disseminate these above-described power sources.

On the other hand, in the ICT market which hastwo sides, consumers and suppliers, platform compe-titions are being developed on a global basis. Theseare attractive on the price side, the supply side, andthe service side. This two-sided market mechanismhas been analysed by using mathematical models(Boudreau and Hagiu, 2009)(Unno and Xu, 2012).In addition, recently, studies regarding real-time dy-namic pricing based on agent modeling and studiesregarding incentive mechanisms (Bacon, 2012) havebeen made.

Smart grid is expected to gain profits from real-time dynamic pricing. This pricing system enablesboth power consumers and power companies to re-flect changes in wholesale prices on the demand side(Samadi et al., 2011). Conversely, auction-basedpower pricing is not an uncommon concept. How-ever, the demand side which participates in auctionsessions is based on renewable energy such as solarenergy. Therefore, electricity generated is very vari-able.

Required studies include those of electricity plat-form design which maximises social welfare whileconsidering the electricity market as a two-sided mar-ket, and those focusing not on a single market, but onmultiple competitive electricity markets. Mechanismdesign in dynamical systems and agent-based gamingmodels are considered to be the best and suitable inorder to optimise participation incentives under suchcircumstances.

Traditional economic models describing changesin markets are less suitable to understand the dynam-ics of interaction in a two-sided market. This is be-cause these models do not account for the emergentprocesses that can happen when multiple actors are

interacting. Agent based simulation is a suitable toolto study the dynamics in markets with many interact-ing actors. When the agent based model is suitable forpolicy makers to experiment with managing the sys-tem, a serious game context can be created to studyboth the impact of decisional strategies as well as thedecision making process of the managers. They canbe confronted with different situation, and it can besystematically studied what type of management andwhich policies are the most effective in guiding sucha transition in the energy market.

An important challenge here is the valid modellingof the population of agents in the model. Realisticagent behaviour is important to make an agent basedgame a tool that provides applicable insights(Jagerand Vegt, 2015).

In two-sided markets with consumers and suppli-ers, platform competitions are being developed on aglobal basis which are attractive on the price side,the supply side, and the service side. This two-sidedmarket mechanism has been analysed by using math-ematical models. However, mathematical modelswere applied to analyse market mechanisms with onlyone or two players(Rochet and Tirol, 2003)(Rochetand Tirol, 2006)(Sannikov, 2008). Therefore, mathe-matical models have limitations in analysing mecha-nisms with multiple diversified players such as con-sumers. In addition, studies regarding ABM-baseddynamic pricing and incentive mechanisms have beenin progress. In these studies, however, the decision-making process of agents was controlled by an algo-rithm. For this reason, there are limitations in thesestudies to analyse complicated decision-making pro-cesses taking into account movements of actual envi-ronments, human behaviour and complex energy con-sumers markets, and corporate management condi-tions. Based on these traditional models, in this re-search, we made an attempt to build a two-sided mar-ket model for electricity markets by applying agent-based gaming.

Serious game sessions have been held in re-cent conferences regarding social simulation (ESSA,2015). As for the traditional approaches of seriousgames, however, societies and environments whichserved as backgrounds were defined by game design-ers. Therefore, they often tend to have determinis-tic characteristics. Real societies, where participat-ing agents are actually thrown into interactions withother agents or a non-linear process, have the prop-erty of complex adaptive systems. Electricity marketsare expected to be such a circumstance as mentionedabove. This requires gaming with an assumption ofcomplex adaptive systems.

Afterward, section 3 describes the research objec-

An Electricity Market Game using Agent-based Gaming Technique for Understanding Energy Transition

315

tives, and section 4 explains the outline of the energyconversion model. Section 5 gives a description ofexperimental environment and section 6 discusses theexperimental results, while section 7 summarises thisresearch.

3 RESEARCH OBJECTIVES

The objective is to analyse what players can obtainmarket ascendancy under what kind of conditions inan electricity market. In order to achieve electricityplatform design which maximises social welfare, thisresearch focuses on aggregators and imbalance ad-justment. Currently, utilisation of market functionsassociated with electricity supply and demand adjust-ment has been considered, with a proposal for estab-lishing a new one-hour-ahead market and a real-timemarket in order for electricity distribution operators,power producers and retailers to procure the most ef-ficient regulated power supplies from these markets(Ministry of Economy, Trade and Industry, 2013).Use of these market prices in imbalance settlement forrenewable energy can secure transparency and fair-ness. This should have positive influence on the ef-ficiency of electricity markets and the promotion ofrenewable energy dissemination (Fig.1).

発電・卸電気事業者

需要家Customers/Consumers

Power Suppliers

小売事業者Aggregators

Figure 1: Imbalance settlement and the electricity market.

Market participants are diverse agents and themarket itself also consists of multiple competitiveplatforms; therefore, these things are considered tobe multi-agent and multi-purpose optimisation prob-lems. Solving such problems requires a multi-agentincentive mechanism, while an appropriate approachis agent-based modelling (ABM). On the other hand,when the decision-making process of power suppli-ers and aggregators is left to machine agents, the al-gorithm’s capability could affect the decision-makingresults. However, a human-agent participatory gam-ing method which has been used for serious games

is more likely to obtain the decision-making resultsthat are close to the actual results when human agentsas players organically connect and consider informa-tion which they obtain from models. In traditional se-rious games, however, environmental changes as thebackground are determined in a deterministic manner.This fact makes it difficult to reproduce the compli-cated movements of an electricity market.

Given that, through this paper, progress has beenmade in our present research based on the follow-ing two points while connecting ABM and seri-ous games and introducing an agent-based gamingmethod which makes it possible to design multi-agentand multi-purpose models.

3.1 Analysis of Market Structure whichBrings about Energy Conversion

System design in electricity markets have a significantinfluence on generation of market rulers. Our addi-tional goal is to design a system which is effectivefor energy conversion to renewable energy. Designof a mechanism for achieving stable electricity sup-ply equilibrium based on utilisation of a wide varietyof energy sources needs to play the role of a platformfor maximising the utility for both electricity suppli-ers and consumers. In order to analyse these struc-tures, we use ABM.

3.2 Comparative Analysis ofDecision-making Structures

While expanding electricity consumers and powerproducers to multiple agents, their behaviour is ex-pressed by using a multi-agent model. With that,we conducted comparative analysis on the decision-making results obtained by introducing participa-tory agent-based gaming. By analysing differencesbrought by each individual agent, we evaluated strate-gies of imbalance adjustment incentives for electric-ity, and government subsidies and tax rate policies. Inaddition, observing the targeted phenomenon not onlyfrom a single viewpoint, but from several differentviewpoints, in order that each phenomenon can be ex-pressed accurately by using only one model (Grimm,2005).

4 ENERGY CONVERSIONGAMING MODEL

In energy conversion gaming models based on agent-based gaming models (Fig.2), in an electricity market

ICAART 2017 - 9th International Conference on Agents and Artificial Intelligence

316

where power producer players and aggregator play-ers participate, power producers make their decisionsbased on electricity sale prices, advertising invest-ments, and plans for power-generation facilities. Saleprices are adjusted based on imbalance settlement insupply and demand with electricity distribution oper-ators.

On the other hand, we can expect that mar-keters, brokers, local public organisations, and non-profit groups which organise electric needs of con-sumers in order to provide energy management ser-vices effectively will participate in electricity mar-kets. They play their roles as aggregators which serveas a bridge between retail players and general house-holds/operators. Aggregators are expected to providea wide variety of services based on advanced energymanagement systems by using smart meters, whiledeveloping demand responses and negawatt1 services.This might allow aggregators to dominate market cir-culation in a two-sided market, and to have the powerto determine not only the price, but also to profit allo-cation. This possibility brings the same structure as ITmarkets including music distribution and smartphoneapp markets, where fierce competition for dominatingmarkets can be caused. Therefore, it is extremely im-portant to study on market system design which canpromote development of renewable energy and soundmarket competition.

The proposed agent-based gaming model is basedon the government plan of energy market reform inJapan(NRE, 2015).In this gaming model, the actualparticipants participate in the game playing the rolesof power producers, electricity retailers, and aggrega-tors. In addition, computer agents also participate inthe market autonomously as a number of consumeragents. The government agents conduct imbalancesettlement based on the predetermined market rules.Based on this gaming model, the game participantscan experience the complexity of this market and theycan design a market system while verifying the effec-tiveness of the system designed. Our ultimate goal isto verify whether real-time characteristics are satisfiedby conducting simulation based on the actual climatedata in order to develop further verification.

4.1 Model Outline

According to the ODD protocol, the section be-low describes the outline of the model. The ODD(Overview, Design concepts, and Details) proto-col was proposed to standardise the published de-scriptions of individual-based and ABMs(Grimm,

1Negawatt power is a theoretical unit of power repre-senting an amount of energy (measured in watts) saved.

2005). The primary objectives of ODD are to makemodel descriptions more understandable and com-plete, thereby making ABMs less subject to criticismfor being irreproducible.

In this model, ’Entities’ are electricity suppliers,aggregators, the government, and consumers. ’Statevariables’ are defined as follows:

• Electricity suppliersSale prices, discount rates for major clients, in-vestments (advertising, thermal, nuclear, and re-newable energy), costs (thermal, nuclear, and re-newable energy), carbon generation rates (ther-mal, nuclear, and renewable energy), power gen-eration amounts (thermal, nuclear, and renewableenergy), operator attractiveness, carbon gas gen-erated, and rate of power failure occurrences

• AggregatorsSale prices, advertising investment, the numberof operators that purchase electricity, and energyproportions (thermal, nuclear, and renewable en-ergy)

• GovernmentImbalance prices, business tax rates, carbon taxrates, and renewable energy investments

• ConsumersNorm effect parameters, information effect pa-rameters, network generation parameters, and thenumber of consumers

’Process overview and scheduling’ are as below.Suppliers generate power, and sell it to consumers andaggregators. While taking into account the environ-ment of consumers and their intentions toward prices,suppliers determine the power generation proportionsof thermal power generation, nuclear power gener-ation, and renewable energy, electricity prices (dis-counts for general/major clients), and advertising in-vestments in order to maximise their own profits. In-crease in the proportion of renewable energy increasesthe power failure probability, resulting in paying theimbalance cost. Additionally, their own competitive-ness declines in proportion to the power failure prob-ability.

Aggregators purchase electricity with discountsfor major clients from suppliers, while re-selling theelectricity to consumers. While taking into accountthe environment of consumers and their intentions to-ward prices, aggregators determine the power gener-ation proportions of thermal power generation, nu-clear power generation, and renewable energy, elec-tricity prices (for general clients), and advertising in-vestments.

While considering their own preferences for elec-tric power and electric power charges, consumers pur-

An Electricity Market Game using Agent-based Gaming Technique for Understanding Energy Transition

317

!""#$"%&'#(

)'*$#(*+',$-%,$(.%#/$&(

Imbalance adjustment

((

)'*$#(-011,2$#(

Supply plan

Discount price

Two-sided market

),%3$#-(

)'*$#(-011,3(1,%3$#(

!""#$"%&'#(1,%3$#(

4'5$#6.$6&(1,%3$#(

(

!0&'6'.'0-(%"$6&(

7'6-0.$#-(

4'5$#6.$6&(

Renewable subsidy Business tax

Carbon tax

)'*$#(#$&%2,$#(82-&#290:'6(1#';$--(

Consumers

Selling price Advertisement

Power plant investment

(Thermal/nuclear/renewable)

Blackout probability

Selling price Advertisement

Balance between

environment and cost

Figure 2: Energy conversion gaming model.

chase electric power from appropriate suppliers. Con-sumers are network-linked with their acquaintancesreceiving the norm effect. The government deter-mines imbalance prices, business taxes, carbon taxes,and renewable energy subsidies. Based on these, thetotal amount of carbon gas generated and the entireprobability of power failure are determined. The goalof the government is to optimise these variables.

Within the game, the degree of market domi-nance by players and decision-making of consumerson neighbouring networks are generated as ’Emer-gence’.

In gaming models, motivations of players aredefined that supplier and aggregator human playersmake their decisions so that they maximise their prof-its while referring to decision-making status of otherparticipants as ’Adaptation’ processes. Alternatively,their attitudes to their own environment could be re-flected. Controlling the amount ofCO2, governmentplayers make their decisions so that tax revenues canbe secured and the power failure probability is main-tained at a lower level. The players in gaming modelsare expected to discuss and learn as a team.

Agent players interact with other suppliers, con-sumers, and the government through the market in thefollowing points. Agent players develop competitionby receiving orders from consumers, establishing theregulations onCO2 emission based on carbon taxeswith the government, implement environmental mea-sures based on renewable subsidies with the govern-

ment, establish the restrictions on renewable energybased on imbalance adjustment prices with the gov-ernment, compete with other companies for attrac-tiveness based on stable electric power supply (powerfailure probability) with consumers, and secure profitsand compete for receiving orders through discountedprices with aggregators.

The initial electric power preferences of con-sumers are stochastically determined in a uniform dis-tribution, while the environmental preferences varydepending on the period. The electric power propor-tions are determined in uniform random numbers ex-pressed by the base proportion +/-10%. Based on thesynthesised attractiveness of prices and electric powerpreferences, suppliers and aggregators are determinedby using roulette selection. The power failure proba-bility is an exponential function based on the renew-able energy proportions.

Decisions made by consumers as realistic agentsare determined based on the norm effect of neighbour-ing market shares on consumer network models(Delreet al., 2007)(Toivonen et al., 2006). As for the normeffect of consumers, the threshold model which is in-fluenced by neighbouring market shares is adopted.Consumers also have information effect functions inwhich they make decision to purchase electricity fromsuitable power suppliers stochastically based on priceand energy sources such as thermal, nuclear and re-newable energy(Kurahashi and Saito, 2013).

ICAART 2017 - 9th International Conference on Agents and Artificial Intelligence

318

5 EXPERIMENT

An electricity gaming model was implemented by theagent programming environments, NetLogo and Hub-Net, which was operated from each terminal con-nected to the local network. Fig.3 shows the screendisplayed for players, while Fig.4 shows that for op-erators.

Figure 3: Player Panel: Networks of consumers and mar-ket shares are graphically observed. Decision-making andmanagement conditions of other players, including the pref-erences of consumers, are observed on a panel. Decision-making and management condition of all players are able tobe observed on a panel in every period.

Figure 4: Operator Panel: A game operator could confirmthe condition of consumers allocated on the network and thecondition of suppliers on the panel.

From any of these screens, the condition of con-sumers allocated on the network and the conditionof suppliers could be confirmed. Consumers wereable to identify order destinations in different colours,so that they could intuitively understand the currentshare condition of suppliers. Electric power prices,investment for power generation facilities, and man-agement information including surplus funds could beconfirmed as supplier conditions. From this playerscreen, each supplier player entered necessary infor-mation such as the electric power prices, advertisinginvestments, investments for thermal power genera-tion, investments for nuclear power generation, in-vestments for renewable energy power generation,and the discount rates for major clients. Aggregate

players determined the price and the energy sourceweight, in addition to the electric power price andadvertising investments, as factors for deciding orderdestinations.

In this experiment, four supplier players, one ag-gregator agent, 500 consumer agents and the govern-ment agent made their decisions for 18 periods. Thethreshold level of norm effect is 0.5. Information ef-fect which indicates price and energy balance of sup-pliers is 0.3. Business tax rate is 35%. Carbon tax rateis 10%. Thermal energy cost is 10 unit / kW, nuclearenergy cost is 5 unit / kW, and renewable energy costis 8 unit / kW, imbalance price is 3 unit / kW, and apreference level between price and energy source ofaggregators is 50%.

The goal of this study is to clarify decisive fac-tors for making decision of energy selection basedon human competitive and collaboration behaviourto be helpful for an incentive design of energy mar-kets. For the purpose, two hypotheses were set inthe experiment. First is that energy transition to re-newable source is achieved by players while keepingtheir profit. Second is that aggregators have abilityto control the energy market through the share of con-sumers’ power market as well as other two-sided mar-kets.

6 RESULTS AND DISCUSSION

The left chart of Fig.5 shows the proportion of eachenergy source, the amount ofCO2 emissions, and thetransition of the power failure probability. In the ini-tial stage, the proportion of thermal power generationexceeded 60%; however, it declined gradually, finallygoing down to less than 40%. This also reduced theamount of carbon emissions (The right chart of Fig.5).The first hypothesis, which energy transition to re-newable source is achieved by players while keepingtheir profit, has been adopted with this result.

On the other hand, the proportions of nuclearpower generation and renewable energy power gen-eration increased. This is because of the influencegiven by the energy orientation of consumers. In par-ticular, the proportion of renewable energy graduallyincreased in tune with the orientation of consumers,while it declined in the later stages. This result mightbe because whereas the power generation proportionof each electric power supplier was inclined towardthe use of thermal power generation in the initialstage, the energy orientation of consumers was about1/3. Therefore, there must have been an incentive thatworked where the order volume increased by chang-ing the power generation investment according to this

An Electricity Market Game using Agent-based Gaming Technique for Understanding Energy Transition

319

!"#"$%&'"

()*'"%+

,-"+.%'

,"+.

!%/012132"'"*/+0*241$"+

5'%*61)/

789

7892".0::01# 5'%*61)/24+1&%&0'0/;

,"+.

Figure 5: Left: Trend of energy source rate, Right: Trend ofCO2 and blackout rate.

!"#$%&'

()%'*&$

+%,%-&.$%

+&/012132#1,4"*%'452%,%'6728'%3%'%,#%

(%'* (%'*

9'130/

:66'%6&/1'2

9$&7%'2;

9$&7%'2<

9$&7%'2=

9$&7%'2>

Figure 6: Left: Trend of consumers’ energy preference, Right: Trend of players’ sales in the power market.

proportion (The left chart of Fig.6).However, the situation, which was originally ex-

pected that the proportion of nuclear power genera-tion decreased, was not observed, while nuclear en-ergy with lower cost and carbon gas emissions con-tinue to be relied on. This result shows that the man-agement of electric power suppliers gave the first pri-ority to maximising their profits, while giving almostno consideration to risks of nuclear power generationaccidents. On the other hand, the aggregator agentmade profit as well as suppliers players, but it couldnot monopolise the electric consumer market becauseone possibility is that the supplier players learnt howto keep their market share in competition from theaggregator(The right chart of Fig.6). The second hy-pothesis, which aggregators have ability to control theenergy market through the share of consumers’ powermarket as well as other two-sided markets, was re-jected with the result.

All of the four players participating in this experi-ment were business people in their 30’s, who mighthave had a custom to make decisions to maximisebusiness profits as corporate managers. They were atthe same time consumers, however, this experimentsuggests that their concepts of accident risks mightsignificantly change when they play a social role as

entities to make corporate decisions.

7 CONCLUSION

In this research, based on agent-based models, seri-ous games, design of electricity market platforms, andsocial network models, we built a model having theitems below as purposes.

1. Feature analysis on electric power imbalance ad-justment for achieving new system designs

2. Design of competitive electricity market plat-forms

3. Design of incentive mechanisms for imbalanceadjustment

4. Evaluation and examination of mechanism designbased on agent-based gaming models

The goal of this study is to clarify decisive fac-tors for making decision of energy selection basedon human competitive and collaboration behaviourto be helpful for an incentive design of energy mar-kets. For the purpose, two hypotheses were set inthe experiment. First is that energy transition to re-newable source is achieved by players while keeping

ICAART 2017 - 9th International Conference on Agents and Artificial Intelligence

320

their profit. Second is that aggregators have abilityto control the energy market through the share of con-sumers’ power market as well as other two-sided mar-kets.

Our experiment confirmed that the energy orien-tation of electric power consumers could give a sig-nificant influence on power generation investment ofelectric power suppliers, and the risk of nuclear en-ergy was underestimated. And the first hypothesiswas adopted and the second was rejected by the exper-iments through the agent-based gaming. These find-ings enabled us to analyse the decision-making pro-cess of people and operators, while being able to ob-tain effective knowledge regarding social ecosystemswhich disseminate renewable energy and adaptive be-haviour.

In the future, we are going to examine combinedmodels with autonomous and human agents to com-pare with them. The autonomous agent-based modelwill show behaviour and attitude as a control group tovalidate the hypotheses more thoroughly. We will alsoconduct several games including autonomous agentsand human players and compare with other modelssuch as an equilibrium model and so on.

REFERENCES

Bacon, D. (2012). Predicting your own effort. volume 2,pages 695–702.

Boudreau, K. and Hagiu, A. (2009).Platform, Markets andInnovation. UK.

Delre, S., Jager, W., and Janssen, M. (2007). Diffusion dy-namics in small-world networks with heterogeneousconsumers.Computational and Mathematical Orga-nization Theory, 13:185–202.

ESSA (2015). http://www.essa2015.org.

Grimm, V. (2005). Pattern-orinted modeling of agent-basedcomplexsystems. Lessons from Ecology, 310:987–991.

Jager, W. and Vegt, G. V. D. (2015).Management of com-plex systems: towards agent based gaming for pol-icy. In Policy practice and digital science: Integratingcomplex systems, social simulation and public admin-istration in policy research, Ed. M. Janssen. SpringerScience in the Public Administration and InformationTechnology series edited Christopher G. Reddick.

Kurahashi, S. and Saito, M. (2013). Informative and nor-mative effects using a selective advertisement.SICEJournal of Control, Measurement, and System Inte-gration, 6(2):76–82.

NRE (2015). Agency for natural resources and energy inJapan.

Rochet, J. and Tirol, J. (2003). Platform competition in two-sided markets.Journal of European Economic Asso-ciation, 1(4):990–1029.

Rochet, J. and Tirol, J. (2006). Two-sided markets :a progress report. Rand Journal of Economics,37(3):645–667.

Samadi, P., Schober, R., and Wong, V. (2011). Optimalenergy consumption scheduling using mechanism de-sign for the future smart grid. pages 369–374.

Sannikov, Y. (2008). A continuous time version of theprincipal-agent problem.Review of Economic Jour-nal Studies, 75(3):957–984.

Toivonen, R., Onnela, J., Saramaki, J., Hyvonen, J., andKaski, K. (2006). A model for social networks. vol-ume 371.

Unno, M. and Xu, H. (2012). Opitmal platform strategiesin the smartphone market.The transactions of the In-stitute of Electrical Engineers of Japan. C, A publica-tion of Electronics, Information and Systems Society,132(3):467–476.

An Electricity Market Game using Agent-based Gaming Technique for Understanding Energy Transition

321