Optimizing Online Auction Bidding Strategies Using Genetic Programming

Size: px
Start display at page:

Download "Optimizing Online Auction Bidding Strategies Using Genetic Programming"

Transcription

1 Optimizing Online Auction Bidding Strategies Using Genetic Programming Ekaterina Smorodkina December 8, 2003 Abstract The research presented in this paper is concerned with creating optimal bidding strategies for online auctions At present time online auctions, that are very popular today, have not been given proper attention The strategies created using genetic programming were tested in an auction simulation environment in order to determine their strength This paper describes the experimental setup, the online auction simulation environment, and the genetic programming approach to developing bidding strategies It also gives a brief introduction to auction theory and genetic programming The last section of the paper discusses possible reasons for why the current setup failed to produce optimal bidding strategies Keywords Auction, English auction, online auction, ebay, genetic programming, evolutionary algorithms, agent, bidding strategy 1 Introduction 11 Overview In the past eight years online auctions have built a large person-to-person trading community on the Internet On any given day, there are millions of items sold through online auctions, such as ebay, Amazon, Yahoo, etc BusinessWeek online [6] reports that in 1999 ebay had 2 million items for sale with gross merchandise sales of $746 million, while Amazon had 16 million items for sale via online auction with gross merchandise sales of $610 million As stated in [1], in 2002 ebay members transacted $1487 billion in annualized gross merchandise sales The popularity of online auctions can be attributed to their ease of use, wide range of categories of items for sale, and availability 24 hours a day, seven days a week However, the biggest attraction of online auction is the possibility to purchase goods for a price significantly lower than the regular store prices Since online auctions play an important role in the trading world today, a better understanding of auction dynamics is necessary 1

2 As can be inferred from [3] and [5], evolutionary algorithms can be used to gain more knowledge about various types of auctions However researchers in computational economics have not given proper attention to online auctions, that are very popular today Online auctions resemble English auctions, also known as the open-outcry auctions or the ascending-price auctions As defined in [2] English auction is a type of sequential auction in which auctioneer directs participants to beat the current, standing bid New bids must increase the current bid by a predefined increment The auction ends when no participant is willing to outbid the current standing bid Then, the participant who placed the current bid is the winner and pays the amount bid In online auctions the current bid is also sequentially incremented However, online auctions differ from English auctions by an explicit time constraint Online auction participants have a limited time during which they can increment the current bid on an item Although English auctions have a time constraint imposed by the auctioneer, it is not as explicit and clear as in the case of online auctions Another major difference between online and English auctions is that the buyers and sellers can participate in online auctions from anywhere in the world, provided they have a computer that is connected to the Internet With English auctions the buyers are restricted to the physical location of where the auction takes place and the sellers typically have representatives that need to be present Finally, in online auctions the buyer is free to determine the amount by which to increment the current bid, as long as this amount is no less than the minimal bid increment imposed by the auction cite, while in English auctions the increment amount is typically fixed As mentioned in [3], the increase in the number of online auctions makes the consumers face the increase in the complexity of computing the right bid Another factor that increases the complexity of the consumer s decision making process is the increase in the number of items he desires to obtain Since most consumers have a restricted amount of money they are able to spend, bidding on multiple items requires a more careful bidding strategy This paper will focus on creating bidding strategies that are optimal for bidding on multiple items and explore the following questions: 1 Is it possible to come up with one all-purpose bidding strategy for various online auction scenarios? 2 How successful is genetic programming (GP) in evolving bidding strategies for online auctions? Various users may have different combinations of items on which they wish to bid, that is why the optimal bidding strategy must be adaptable to the various users requests However, since there are millions of auction listings for thousands of categories, it is safe to assume that the state of the auction, which is characterized by the types and the number of items listed, remains static Subsections 12, 13, and 14 give an overview of auction theory, computer agents, and genetic programming respectively Section 2 summarizes some previous work that has been done in creating autonomous bidding agents Section 3 2

3 describes the auction simulation used for this experiment Section 4 explains the methodology behind creating the bidding strategies Section 5 describes the experiment designed to test the strategies and presents the results of this research Finally, Section 6 discusses the conclusions derived from the results of the experiment and explores the possible extentions of this research 12 Auction Theory This section gives an overview of how auction mechanisms work It introduces various auction types and discusses bid formulation as a strategic decision Since online auctions have a strong resemlence with English auctions, bidding in English auctions is explored in more detail Simply stated, an auction can be thought of as a marketplace, where a seller tries to obtain as much money as possible and a buyer wants to pay as little as possible Typically an auction is used when a seller is unsure about the market values of the goods he wants to sell, while a buyer has a good idea of what he is willing to pay The amount a buyer is willing to pay may change depending on competitors reactions As mentioned in [4], the major difference between traditional markets and auction markets is that in the former the price is determined strictly by supply and demand, while in the later the trading rules influence the price and allocations Potential buyers compete with each other by submitting bids high enough to win an auction Sellers submit an ask price based on their cost of supplying the good The lack of complete information about buyers valuations and sellers costs results in a difference between ask and bid prices The bidding proces usually discovers the common bid and ask prices that equate supply and demand [4] Thus in an auction the price is set not by the seller but by the buyer However the seller sets the rules by choosing the type of auction to be used In [4] the authors define four basic auction types: English, Dutch, first-price sealed-bid, and Vickrey English auctions have the ascending outcry format The price is successively raised until one bidder remains The good is sold to the last remaining bidder at a price just above that which sees the second last bidder retire Dutch auctions are the reverse of English auctions, with bids announced in a descending order A bidder wins by being the first to accept an announced bid and pays that price First-price sealed-bid auctions require bidders to submit single confidential bids to the seller The bidder with the highest bid wins and pays that bid Vickrey auctions have a second-price sealed-bid format The bidder making the highest bid wins and pays the next highest bid There are two reasons for buyers to participate in an auction The first reason is a desire to purchase goods for personal consumption and the second reason 3

4 is to acquire goods for commercial use or resale In the first case a buyer has a private-valuation of goods and he is willing to pay up to a certain maximum, independent of valuations of other buyers Buying art for personal pleasure is an example of private-valuation In the second case the goods are worth the same amount to all buyers but this amount is unknown Each buyer estimates the value of such goods using the same measurements This is known as a common-value assumption With the common-value assumption bid formation largely depends on competitors reactions In an English auction the buyers continue to bid so long as the current bid price remains below their valuation of the good (either private-valuation or common-valuation) A buyer s private-valuation typically remains the same throughout the auction and a buyer withdraws from the auction if the current bid price exceeds this value A common-value may change during the auction as the buyers evaluate their competitors bids Lack of information about the actual worth of goods and lack of bidding experience may lead to what is known as winner s curse Winner s curse is the situation when the winner of an auction pays more than what the good is actually worth Thus the auction winner looses because his profits decrease 13 Agents Before going any further with the description of this research, a notion of agents needs to be introduced, since agents are a crucial part of the experiment A simple definition of an agent given in [9] is an agent is just something that acts However, simply creating just something that acts can hardly solve any problem This research is concerned with creating rational agents, the agents that produce the best outcome Computer agents have special attributes and qualities that distinguish them from other computer programs An agent operates within an environment and must be able to perceive the environment and act upon it Agent s actions depend on the sequence of inputs he receives from the environment The behavior of an agent can be described by the agent function that maps any given sequence of inputs to an action An agent is typically aware of his own actions, however he does not necesseraly know about the effect of his actions on the environment In this research the buyers that are participating in the auction are represented by agents The environment in which the agents operate is modeled using an auction simulation described in Section 3 An agent s bidding strategy is the agent function Throughout the experiment agents become more rational and learn to produce better outcomes through the evolution of their strategies 14 Genetic Programming Problem representation is a key issue in deciding which evolutionary algorithm to use The representation scheme chosen for the problem defines the window through which the system observes the world [8] For example with the fixed length strings of genetic algorithms only a certain set of parameters or other 4

5 items can be represented Genetic algorithms with variable length strings are capable of representing if-then rules Clissifier systems increase the complexity of string-based if-then rules Genetic programming paradigm is powerful enough to represent entire computer programs The individuals in genetic programming are compositions of functions and terminals appropriate to the particular problem domain [8] A set of functions can include mathematical functions, arithmetic operations, conditional and logical operations, and program specific functions Each function in the set must be defined for all values in the range of all other functions The search space of genetic programming is the hyperspace of all possible combinations of functions and terminals A convenient way to manipulate the compositions of functions and terminals is through the use of S-expressions represented in a tree data structure These expressions are executed and are assigned a fitness that depends on the outcomes of the execution The typical genetic operations of genetic programming are recombination and mutation A specific type of recombination, subtree crossover, gives birth to new individuals with a potentially better fitness This process is described in [7] As shown in Figure 1, to perform subtree crossover two parent trees are selected A subtree is chosen in Parent 1 and in Parent 2 The child tree is produced by replacing the subtree of Parent 1 with the subtree of Parent 2 As a result the child tree contains characteristics of the parent trees and has the same constraints on the number of inputs into each operator as the parent trees Mutation in genetic programming is performed by randomly selecting a tree node to mutate and changing its value If a function node is selected it is changed to a different function In the case of mutating a function node the number of function inputs must be verified against the number of child nodes If a terminal node is selected for mutation it is changed to a different terminal In this research genetic programming paradigm is applied to come up with successfull online auction bidding strategies 2 Previous Work Some work has already been done in developing bidding agents for online auctions Anthony and Jennings [3] developed autonomous agents that participate in multiple auctions These agents monitor and collect information from multiple ongoing auctions, and use this information to determine the amount of the next bid A genetic algorithm (GA) was used to codify the bidding strategies The authors took into consideration the consumer s private valuation for each item and the consumer s desire for bargain One drawback of this work is the lack of dynamic environment, since the evolution of agents is performed in an offline fashion Ashlock [5] used a more compex strategy representation, GP-Automata, to evolve bidding strategies for the electricity marketplace GP-Automata combie finite state automata with genetic programming The strategies developed 5

6 Parent 1 > Parent 2 > * + - % p1 p2 p3 p4 p1 * p3 p4 Child > p2 p3 * % p1 p2 p3 p4 Figure 1: Subtree Crossover in this research are able to adapt to the changing environment Although the results of the research are specific to the electricity marketplace, they are applicable for other markets The work previously done in this field shows that evolutionary algorithms are an appropriate approach to developing bidding strategies The approache involving genetic programming proved to be successful in dynamic environments, while GA s was more useful in an stable marketplace 3 Auction Simulation In this experiment we are concerned with the bidding strategies for online auctions Almost all online auctions have the same features Since ebay is one of the most popular online auctions, an auction simulator, ebaysim, was created based on the example of ebay ebaysim operates in a matter similar to an English auction where the price on each item is raised successively until the item listing closes In ebaysim each item is listed for a certain length of time When this time expires the listing closes and the agent who has the highest bid on the item is required to pay the amount of the winning bid Each item to be auctioned off starts with a minimum bid amount and contains information about the retail price on this item Table 1 demonstrates an example of four item listings at the start of the auction, and Table 2 shows the same item listings at the end of the auction Note that item 1 is listed twice If the same item is listed multiple times, each one of those listings must specify the the same retail price, while other parameters may differ At the start of an auction no agents have submitted any bids, so the Bid on each item is zero and the Agent ID is empty At the end of the auction in this example Agent 10 has the highest bid 6

7 on the first instance of item 1 and on item 2, Agent 5 has the highest bid on the second instance of item 1, and Agent 9 holds the highest bid on item 3 These agents buy the items by paying the amount of their highest bid ID Retail Price Time Bid Start Price Agent ID 1 $ units $0 $500 none 1 $ units $0 $700 none 2 $ units $0 $950 none 3 $ units $0 $400 none Table 1: Four Item Listings at the Start of an Auction ID Retail Price Time Bid Start Price Agent ID 1 $ units $1275 $ $ units $1300 $ $ units $1960 $ $ units $1525 $400 9 Table 2: Four Item Listings at the End of an Auction The agents participating in the auction have a shopping list of items that they need to purchase When the auction begins the agents start submitting their bids on the items on their shopping lists Bid calculations and submittions are performed concurrently and asynchronously This means that while one agent is placing a bid on an item another agent may be calculating his next bid on that same item When an agent looses the status of the highest bidder on an item, he receives a notification about it If the maximum time for auctioning off an item is reached, no more bids can be placed on this item Figure 2 demonstrates one possible scenario of concurrent bid submission process Although the agents know the retail price of each item they are still able to get a winner s curse - to bid higher than the retail price of the item However a desired strategy should not result in a winner s curse There may be multiple instances of the same item for sale during the simulation However, the agent needs to make sure that he only bids on one instance of the item at a time 4 Methodology 41 Agent Design An agent in this experiment is designed to be able to bid on multiple auction items at the same time These items are listed in the agent s shopping list An agent enters the auction house with a fixed amount of money that he can 7

8 Agent 1 Item 1 Agent 1 Item 1 Agent 2 Agent i Agent K Item j Item M Agent 2 Agent i Agent K Item j Item M (a) Solid arrows point to the agents that currently have the higheset bid on the item (b) Dashed arrows show which agents are submitting the bids Agent 1 Item 1 Agent 2 Agent i notify Item j Agent K item listing closed Item M (c) Dotted arrows show the message sent to the agents Now Agent 2 is the highest bidder on Item j Figure 2: Bid Submission Process 8

9 spend, this amount is referred to as account balance The sum of the agent s bids cannot exceed his account balance An agent knows on which items he currently has the highest bid If an agent looses the status of the highest bidder, he is immediately notified about this change An agent calculates his bids using a bidding strategy, the details of which are explained in Subsection 42 Figure 3 shows an example agent at the start of an auction Here the agent has not yet had a chance to place his bids, thus the bid on each item in his list is zero Agent 10: $40 Shopping List Item 1 Item 2 Item 5 Item 9 Bid List Item 1: $0 Item 2: $0 Item 5: $0 Item 9: $0 Figure 3: Agent 10 at the Start of an Auction Figure 4 shows the same agent at the end of the auction Now this agent holds the highest bids of items 1 and 2 As agents submit their bids, the bid amount is subtracted from their balances When an agent looses the status of the highest bidder on an item, his balance is incremented by the amount of his bid on that item, since he no longer will be required to buy the item Agent 10: $1775 Shopping List Item 1 Item 2 Item 5 Item 9 Bid List Item 1: $1275 Item 2: $ 1960 Item 5: $0 Item 9: $0 Figure 4: Agent 10 at the End of an Auction The agents proceed submitting their bids until all the item listings close The agents goal is to obtain all the items on their lists at the lowest price 9

10 possible 42 Strategy Representation An agent s bidding strategy is represented by an expression tree The expression takes in a fixed number of input parameters and produces the next bid on an item After a careful analysis of the factors that humans consider when bidding in an auction the following parameters were chosen as inputs to the expression tree: 1 Account balance - the amount of money an agent has 2 Item s retail price - the store price of the listed item 3 The total number of items in the agent s list - the size of the shopping list 4 The number of items on the list on which the agent does not have the highest bid - the number of items on the shopping list that the agent still needs to buy 5 The higest bid among all the instances of this item - if there are several listings with the same item id, this is the highest bid among all those listings 6 The lowest bid among all the instances of this item - if there are several listings with the same item id, this is the lowest bid among all those listings 7 The current bid on this item listing 8 A coefficient c in the range (0,1) - this is an additional parameter that may be necessary to adjust the bid It must be noted that this list may be incomplete - there may be other factors that need to be considered in the calculation of the next bid However this list gives a starting point for developing smart bidding strategies The factors above are the input parameters to the expression trees, or strategies, that act as equations for calculating a bid Some of the initial strategies are generated randomly while others are predefined as intuitive to the author Figure 5 shows an example of a predefined strategy In this example the current bid, represented by the right sub-tree, is incremented by a small amount, defined in the left sub-tree Note that the strategy in Figure 5 uses only four input parameters It is assumed that during the evolution process other parameters will be added to the strategy To simplify the task, the operators used by the expression trees are restricted to binary This simplifies the implementation of recombination and mutation by eliminating the need to check that the operators and the number of inputs into them match These operators are: addition, subtraction, division, multiplication, modulus, max, and min The list of operators can be expanded in the future 10

11 + p1 / * p4 p2 p3 Figure 5: Predefined Strategy Example: p1 - current bid, p2 - balance, p3 - coefficient, p4 - the number of items in the agent s list on which he does not have the highest bid 43 Fitness Evaluation Agents fitness is evaluated at the end of the auction round (ie, when all the item listings are closed) In an experiment of this sort fitness evaluation is a subjective matter and depends on what the human user values the most Some users may emphasize the importance of winning an item in the auction, others would consider it more valuable to receive a big discount on an item Since this experiment is concerned with a general bidding strategy, a combination of factors must be considered when evaluating an agent s fitness Let N be the number of items an agent obtains from the auction, M be the total number of items on the agent s list, and Max be the length of the longest shopping list among all agents Then below is a list of criteria that are used in computing an agent s fitness and that several users may find important 1 Relative discount received (Relative Discount = N i=1 (retail pricei highest bidi) N i=1 retail price i 2 The agent s ability to maximize the relative number of items obtained (Relative Number of Items Obtained = N M Max ) The fitness of each agent is calculated using the fitness function shown in Equation 1 The parameter Max in the denominator is necessary to prevent partiality towards the agents with short shopping lists This fitness function rewards the agents that obtain more items on their list at a smaller price Since agents are allowed to place bids that are higher than the retail price of the item on which they are bidding, it is possible to have a negative fitness Allowing bids higher than the item s retail price should help detect agents with undesired strategies, since their fitness will be negative ) N N M Max i=1 (retail price i highest bid i ) N i=1 retail price i (1) 11

12 For an example of fitness evaluation consider the agent shown in Figure 4 In this case N = 2, M = 4, N i=1 (retail price i highest bid i ) = 785, and N i=1 retail price i = 405 Let Max = 30, then the fitness assigned to this 2 agent is = Note that with Max = 30 no fitness can be greater than Agent Reproduction At the end of each auction round a new population of agents is produced The new agents are variations of the agents in the previous population These variations are produced by means of tree recombination and mutation as described in Section 14 The parents that are used to produce a new generation of agents are chosen using proportional selection Scaling of proportional selection is unnecessary in this case This can be justified by the fitness function: the fitness is a product of the relative discount received and the relative number of items obtained A 1 fitness can have a value in the range of (-X, Max ), where Max is the length of the longest shopping list among all agents, and X is an unknown positive real number X is unknown because the agents are allowed to place bids as high as they want Thus good and bad strategies will have significanly different fitnesses in this range Since negative fitness values are allowed it is possible that the total population fitness will be negative or will equal to zero Therefore the proportional selection must be adjusted as shown in Table 3 1 Total Population Fitness Agent s Fitness Selection Chance AgentF itness > 0 0 T otalf itness > 0 < > 0 AgentF itness AgentF itness < T otalf itness = 0 < 0 1 AgentF itness = 0 = 0 0 Table 3: Parent Selection Chance Mutation is an important genetic operator when it comes to producing new variations of strategies A random mutation may change a strategy that is close to an optimal solution into one that is optimal However, mutations may also be harmful and result in the decrease of the solution s optimality When too much mutation is introduced, the search for the optimal solution is no longer a directed search but rather a random one Therefore mutation chance must be carefully considered Since both genetic operators: crossover and mutation, result in drastic differences between the parents and the children, only one of them is used at each generation This approach avoids too much destruction of the parent strategies, which would lead the algorithm to a random search 1 The TotalFitness is defined as pop size i=1 fitness(agent i ) 12

13 45 Competing for Survival The strategies are passed on to the next generation using the Elitist strategy If a child strategy results in a better fitness than some strategy in the current population, it is passed on to the next generation, while a weaker strategy is removed from the original population Due to the dynamic nature of this experiment a strategy that produces a high fitness in one auction round may be a weak strategy under a different auction setup 5 Experiments and Results Agents trial strategies were tested in the simulation environment described in Section 3 It would be very inefficient and resource consuming to have millions of item listings in the simulation, the way real online auctions do Thus it was necessary to introduce modifications to the auction state (ie, the number of items listed and their characteristics) at the start of each auction round These modifications ensure that over time the simulation environment has significant resemblence with real online auctions The following environment characteristics were changed throughout the experiment: 1 Number and types of items in the auction 2 The starting bid amount on each item 3 Bidding time left for each item in the auction The agents shopping lists were also modified before the start of each auction round This ensured that if good strategies are to be produced throughout the experiment, they can be applied to any combination of items on the shopping list, given a reasonable account balance The success of the experiment was measured by the increase in the average population fitness After running the experiment through 67,000 generations the average fitness increased from to , which is an 11% increase However, the best strategy at generation 67,000 was one of the pre-defined strategies: current bid + c, where c = So after 67,000 generations the genetic programming approach failed to produce a new effective bidding strategy 6 Discussion and Conclusions Since the experiment failed to produce any significant results, it is not possible to conclude that genetic programming approach is appropriate for evolving bidding strategies for online auctions However, it is still too early to throw away this idea The failure to produce any good results may be attributed to several factors: 13

14 1 The agents used in the experiment lacked initial knowledge of auction theory The pre-defined strategies were unable to provide enough information about bidding in online auctions 2 Although it is necessary to have a dynamic environment in simulating an online auction, the simulation created for this experiment may have been more dynamic than what genetic programming could handle Making the environment more static towards the beginning of the experiment may have increased the overall population fitness 3 The agents designed for this experiment did not keep track of the bidding history The bidding history may be a crucial element in calculating the next bid Implementing bidding history in the future may result in a more successful experiment 4 The bidding strategies were not evolved until the end of the auction Modifying bad strategies during the auction may improve the overall performance Although the genetic programming approach to this problem did not prove to be successful this time, further experiments need to be performed in search for the optimal bidding strategy The future experiments will be modified to reflect the four failure factors listed above References [1] [2] htmlbr [3] Patricia Anthony and Nicholas R Jennings Evolving bidding strategies for multiple auctions [4] Chris Chan, Patrick Laplagne, and David Appels The role of auctions in allocating public resources php, 2003 [5] DAshlock, CW Richter Jr, and GBSheble Comprehensive bidding strategies with genetic programming/finite state automata In Power Systems, IEEE Transactions, pages , November 1999 [6] Robert D Hof and Linda Himelstein /99_22/b htm [7] Jr Kenneth E Kinnear Derivative methods in genetic programming In Evolutionary Computation 1, pages , Philadelphia, PA 19106, USA, 2000 Institute of Physics Publishing (IOP Publishing Ltd) 14

15 [8] John R Koza Genetically breeding populations of computer programs to solve problems in artificial intelligence In Proceedings of the Second International Conference on Tools for AI, Herndon, Virginia, USA, pages IEEE Computer Society Press, Los Alamitos, CA, USA, [9] Stuart Russell and Peter Norvig Artificial intelligence: A modern approach,

SELLER AGENT FOR ONLINE AUCTIONS

SELLER AGENT FOR ONLINE AUCTIONS SELLER AGENT FOR ONLINE AUCTIONS P. Anthony School of Engineering and Information Technology, Universiti Malaysia Sabah Locked Bag 2073,88999 Kota Kinabalu Sabah, Malaysia J. A. Dargham School of Engineering

More information

A Heuristic Bidding Strategy for Multiple Heterogeneous Auctions

A Heuristic Bidding Strategy for Multiple Heterogeneous Auctions Patricia Anthony & Nicholas R. Jennings Dept. of Electronics and Computer Science University of Southampton, UK from International Conference on Electronic Commerce (ICEC) 2003, Pittsburgh, PA presented

More information

Software Frameworks for Advanced Procurement Auction Markets

Software Frameworks for Advanced Procurement Auction Markets Software Frameworks for Advanced Procurement Auction Markets Martin Bichler and Jayant R. Kalagnanam Department of Informatics, Technische Universität München, Munich, Germany IBM T. J. Watson Research

More information

Introduction to Artificial Intelligence. Prof. Inkyu Moon Dept. of Robotics Engineering, DGIST

Introduction to Artificial Intelligence. Prof. Inkyu Moon Dept. of Robotics Engineering, DGIST Introduction to Artificial Intelligence Prof. Inkyu Moon Dept. of Robotics Engineering, DGIST Chapter 9 Evolutionary Computation Introduction Intelligence can be defined as the capability of a system to

More information

Difinition of E-marketplace. E-Marketplaces. Advantages of E-marketplaces

Difinition of E-marketplace. E-Marketplaces. Advantages of E-marketplaces Difinition of E-marketplace a location on the Internet where companies can obtain or disseminate information, engage in transactions, or work together in some way. E-Marketplaces Assist. Prof. Dr. Özge

More information

Traditional auctions such as the English SOFTWARE FRAMEWORKS FOR ADVANCED PROCUREMENT

Traditional auctions such as the English SOFTWARE FRAMEWORKS FOR ADVANCED PROCUREMENT SOFTWARE FRAMEWORKS FOR ADVANCED PROCUREMENT A range of versatile auction formats are coming that allow more flexibility in specifying demand and supply. Traditional auctions such as the English and first-price

More information

What is Evolutionary Computation? Genetic Algorithms. Components of Evolutionary Computing. The Argument. When changes occur...

What is Evolutionary Computation? Genetic Algorithms. Components of Evolutionary Computing. The Argument. When changes occur... What is Evolutionary Computation? Genetic Algorithms Russell & Norvig, Cha. 4.3 An abstraction from the theory of biological evolution that is used to create optimization procedures or methodologies, usually

More information

EVOLUTIONARY ALGORITHMS AT CHOICE: FROM GA TO GP EVOLŪCIJAS ALGORITMI PĒC IZVĒLES: NO GA UZ GP

EVOLUTIONARY ALGORITHMS AT CHOICE: FROM GA TO GP EVOLŪCIJAS ALGORITMI PĒC IZVĒLES: NO GA UZ GP ISSN 1691-5402 ISBN 978-9984-44-028-6 Environment. Technology. Resources Proceedings of the 7 th International Scientific and Practical Conference. Volume I1 Rēzeknes Augstskola, Rēzekne, RA Izdevniecība,

More information

Evolving Bidding Strategies for Multiple Auctions

Evolving Bidding Strategies for Multiple Auctions Evolving Bidding Strategies for Multiple Auctions Patricia Anthony and Nicholas R. Jennings 1 Abstract. Due to the proliferation of online auctions, there is an increasing need to monitor and bid in multiple

More information

A Heuristic Bidding Strategy for Multiple Heterogeneous Auctions

A Heuristic Bidding Strategy for Multiple Heterogeneous Auctions A Heuristic Bidding Strategy for Multiple Heterogeneous Auctions Patricia Anthony Dept. of Electronics and Computer Science University of Southampton Highfield, Southampton, SO17 1BJ, UK. pa99r@ecs.soton.ac.uk

More information

Outline. Part I - Definition of the Negotiation Process. Negotiation. Protocols, Strategies and Architectures for Automated Negotiation

Outline. Part I - Definition of the Negotiation Process. Negotiation. Protocols, Strategies and Architectures for Automated Negotiation Protocols, Strategies and Architectures for Automated Negotiation Claudio Bartolini HP Labs Bristol Bologna, November 17, 2000 Outline Part I Definition of the Negotiation Process Part II Importance of

More information

Evolutionary Computation

Evolutionary Computation Evolutionary Computation Evolution and Intelligent Besides learning ability, intelligence can also be defined as the capability of a system to adapt its behaviour to ever changing environment. Evolutionary

More information

Mysimon.com. Dynamic Pricing. The Internet Changes Pricing Strategies 11/30/2011 PRICING IN DIGITIAL MARKETING

Mysimon.com. Dynamic Pricing. The Internet Changes Pricing Strategies 11/30/2011 PRICING IN DIGITIAL MARKETING PRICING IN DIGITIAL MARKETING Assist. Prof. Dr. Ozge Ozgen 11-5 The Internet Changes Pricing Strategies Price is the sum of all values that buyers exchange for the benefits of a good or service. Throughout

More information

Genetic Algorithm based bargaining agent for Implementing Dynamic Pricing on Internet

Genetic Algorithm based bargaining agent for Implementing Dynamic Pricing on Internet Genetic Algorithm based bargaining agent for Implementing Dynamic Pricing on Internet Kumar Ujjwal MS- AI University of Georgia, USA Jay Aronson Terry College of Business, University of Georgia, USA Abstract

More information

GENETIC ALGORITHMS. Narra Priyanka. K.Naga Sowjanya. Vasavi College of Engineering. Ibrahimbahg,Hyderabad.

GENETIC ALGORITHMS. Narra Priyanka. K.Naga Sowjanya. Vasavi College of Engineering. Ibrahimbahg,Hyderabad. GENETIC ALGORITHMS Narra Priyanka K.Naga Sowjanya Vasavi College of Engineering. Ibrahimbahg,Hyderabad mynameissowji@yahoo.com priyankanarra@yahoo.com Abstract Genetic algorithms are a part of evolutionary

More information

Measuring the Benefits to Sniping on ebay: Evidence from a Field Experiment

Measuring the Benefits to Sniping on ebay: Evidence from a Field Experiment Measuring the Benefits to Sniping on ebay: Evidence from a Field Experiment Sean Gray New York University David Reiley 1 University of Arizona This Version: September 2004 First Version: April 2004 Preliminary

More information

Strategic bidding for multiple units in simultaneous and sequential auctions

Strategic bidding for multiple units in simultaneous and sequential auctions Strategic bidding for multiple units in simultaneous and sequential auctions Stéphane Airiau, Sandip Sen & Grégoire Richard Mathematical & Computer Sciences Department University of Tulsa 600 S. College

More information

Developing a Bidding Agent for Multiple Heterogeneous Auctions

Developing a Bidding Agent for Multiple Heterogeneous Auctions Developing a Bidding Agent for Multiple Heterogeneous Auctions PATRICIA ANTHONY and NICHOLAS R. JENNINGS University of Southampton Due to the proliferation of online auctions, there is an increasing need

More information

Sponsored Search Markets

Sponsored Search Markets COMP323 Introduction to Computational Game Theory Sponsored Search Markets Paul G. Spirakis Department of Computer Science University of Liverpool Paul G. Spirakis (U. Liverpool) Sponsored Search Markets

More information

A Simulation-Based Model for Final Price Prediction in Online Auctions

A Simulation-Based Model for Final Price Prediction in Online Auctions 經濟與管理論叢 (Journal of Economics and Management), 2007, Vol. 3, No. 1, 1-16 A Simulation-Based Model for Final Price Prediction in Online Auctions Shihyu Chou, Chin-Shien Lin, Chi-hong Chen, Tai-Ru Ho, and

More information

Competitive Markets. Jeffrey Ely. January 13, This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License.

Competitive Markets. Jeffrey Ely. January 13, This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License. January 13, 2010 This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License. Profit Maximizing Auctions Last time we saw that a profit maximizing seller will choose

More information

Generational and steady state genetic algorithms for generator maintenance scheduling problems

Generational and steady state genetic algorithms for generator maintenance scheduling problems Generational and steady state genetic algorithms for generator maintenance scheduling problems Item Type Conference paper Authors Dahal, Keshav P.; McDonald, J.R. Citation Dahal, K. P. and McDonald, J.

More information

Generative Models for Networks and Applications to E-Commerce

Generative Models for Networks and Applications to E-Commerce Generative Models for Networks and Applications to E-Commerce Patrick J. Wolfe (with David C. Parkes and R. Kang-Xing Jin) Division of Engineering and Applied Sciences Department of Statistics Harvard

More information

Diffusion Mechanism Design

Diffusion Mechanism Design 1 / 24 Diffusion Mechanism Design Dengji Zhao ShanghaiTech University, Shanghai, China Decision Making Workshop @ Toulouse 2 / 24 What is Mechanism Design What is Mechanism Design? What is Mechanism Design

More information

Computational Intelligence Lecture 20:Intorcution to Genetic Algorithm

Computational Intelligence Lecture 20:Intorcution to Genetic Algorithm Computational Intelligence Lecture 20:Intorcution to Genetic Algorithm Farzaneh Abdollahi Department of Electrical Engineering Amirkabir University of Technology Fall 2012 Farzaneh Abdollahi Computational

More information

Genetic Algorithms for Optimizations

Genetic Algorithms for Optimizations Genetic Algorithms for Optimizations 1. Introduction Genetic Algorithms (GAs) are developed to mimic some of the processes observed in natural evolution. GAs use the concept of Darwin's theory of evolution

More information

bidding for multiple units in simultaneous and sequential auctions.

bidding for multiple units in simultaneous and sequential auctions. From: AAAI Technical Report WS-02-10. Compilation copyright 2002, AAAI (www.aaai.org). All rights reserved. Strategic bidding for multiple units in simultaneous and sequential auctions St@phane Airiau

More information

TIMETABLING EXPERIMENTS USING GENETIC ALGORITHMS. Liviu Lalescu, Costin Badica

TIMETABLING EXPERIMENTS USING GENETIC ALGORITHMS. Liviu Lalescu, Costin Badica TIMETABLING EXPERIMENTS USING GENETIC ALGORITHMS Liviu Lalescu, Costin Badica University of Craiova, Faculty of Control, Computers and Electronics Software Engineering Department, str.tehnicii, 5, Craiova,

More information

The Ascending Bid Auction Experiment:

The Ascending Bid Auction Experiment: The Ascending Bid Auction Experiment: This is an experiment in the economics of decision making. The instructions are simple, and if you follow them carefully and make good decisions, you may earn a considerable

More information

Optimisation and Operations Research

Optimisation and Operations Research Optimisation and Operations Research Lecture 17: Genetic Algorithms and Evolutionary Computing Matthew Roughan http://www.maths.adelaide.edu.au/matthew.roughan/ Lecture_notes/OORII/

More information

Auction Theory An Intrroduction into Mechanism Design. Dirk Bergemann

Auction Theory An Intrroduction into Mechanism Design. Dirk Bergemann Auction Theory An Intrroduction into Mechanism Design Dirk Bergemann Mechanism Design game theory: take the rules as given, analyze outcomes mechanism design: what kind of rules should be employed abstract

More information

Genetic Algorithm: An Optimization Technique Concept

Genetic Algorithm: An Optimization Technique Concept Genetic Algorithm: An Optimization Technique Concept 1 Uma Anand, 2 Chain Singh 1 Student M.Tech (3 rd sem) Department of Computer Science Engineering Dronacharya College of Engineering, Gurgaon-123506,

More information

Multiagent Resource Allocation 1

Multiagent Resource Allocation 1 Multiagent Resource Allocation 1 Michal Jakob Agent Technology Center, Dept. of Computer Science and Engineering, FEE, Czech Technical University AE4M36MAS Autumn 2014 - Lecture 11 Where are We? Agent

More information

Part 1: Motivation, Basic Concepts, Algorithms

Part 1: Motivation, Basic Concepts, Algorithms Part 1: Motivation, Basic Concepts, Algorithms 1 Review of Biological Evolution Evolution is a long time scale process that changes a population of an organism by generating better offspring through reproduction.

More information

Combinatorial Auctions

Combinatorial Auctions T-79.7003 Research Course in Theoretical Computer Science Phase Transitions in Optimisation Problems October 16th, 2007 Combinatorial Auctions Olli Ahonen 1 Introduction Auctions are a central part of

More information

Game Theory: Spring 2017

Game Theory: Spring 2017 Game Theory: Spring 2017 Ulle Endriss Institute for Logic, Language and Computation University of Amsterdam Ulle Endriss 1 Plan for Today This and the next lecture are going to be about mechanism design,

More information

10. Lecture Stochastic Optimization

10. Lecture Stochastic Optimization Soft Control (AT 3, RMA) 10. Lecture Stochastic Optimization Genetic Algorithms 10. Structure of the lecture 1. Soft control: the definition and limitations, basics of epert" systems 2. Knowledge representation

More information

Selecting an Optimal Compound of a University Research Team by Using Genetic Algorithms

Selecting an Optimal Compound of a University Research Team by Using Genetic Algorithms Selecting an Optimal Compound of a University Research Team by Using Genetic Algorithms Florentina Alina Chircu 1 (1) Department of Informatics, Petroleum Gas University of Ploiesti, Romania E-mail: chircu_florentina@yahoo.com

More information

Market Based Control of SCP Congestion in Intelligent Networks

Market Based Control of SCP Congestion in Intelligent Networks Market Based Control of SCP Congestion in Intelligent Networks Frikkie Scholtz 1 and Hu Hanrahan 2 1 Telkom SA Ltd., Private Bag x74, Pretoria 0001, e-mail: scholtfj@telkom.co.za 2 University of the Witwatersrand,

More information

Chapter Fourteen. Topics. Game Theory. An Overview of Game Theory. Static Games. Dynamic Games. Auctions.

Chapter Fourteen. Topics. Game Theory. An Overview of Game Theory. Static Games. Dynamic Games. Auctions. Chapter Fourteen Game Theory Topics An Overview of Game Theory. Static Games. Dynamic Games. Auctions. 2009 Pearson Addison-Wesley. All rights reserved. 14-2 Game Theory Game theory - a set of tools that

More information

Genetic algorithms. History

Genetic algorithms. History Genetic algorithms History Idea of evolutionary computing was introduced in the 1960s by I. Rechenberg in his work "Evolution strategies" (Evolutionsstrategie in original). His idea was then developed

More information

What is Genetic Programming(GP)?

What is Genetic Programming(GP)? Agenda What is Genetic Programming? Background/History. Why Genetic Programming? How Genetic Principles are Applied. Examples of Genetic Programs. Future of Genetic Programming. What is Genetic Programming(GP)?

More information

PARALLEL LINE AND MACHINE JOB SCHEDULING USING GENETIC ALGORITHM

PARALLEL LINE AND MACHINE JOB SCHEDULING USING GENETIC ALGORITHM PARALLEL LINE AND MACHINE JOB SCHEDULING USING GENETIC ALGORITHM Dr.V.Selvi Assistant Professor, Department of Computer Science Mother Teresa women s University Kodaikanal. Tamilnadu,India. Abstract -

More information

Domain Auction and Priority Checking Analysis

Domain Auction and Priority Checking Analysis Domain Auction and Priority Checking Analysis Subhashini.V 1, Babu.M 2 1 M.E (CSE), Department of CSE, G.K.M College of Engineering, Chennai, TamilNadu, India 2 Ph.D Research Scholar, Department of CSE,

More information

Lecture 7 - Auctions and Mechanism Design

Lecture 7 - Auctions and Mechanism Design CS 599: Algorithmic Game Theory Oct 6., 2010 Lecture 7 - Auctions and Mechanism Design Prof. Shang-hua Teng Scribes: Tomer Levinboim and Scott Alfeld An Illustrative Example We begin with a specific example

More information

VISHVESHWARAIAH TECHNOLOGICAL UNIVERSITY S.D.M COLLEGE OF ENGINEERING AND TECHNOLOGY. A seminar report on GENETIC ALGORITHMS.

VISHVESHWARAIAH TECHNOLOGICAL UNIVERSITY S.D.M COLLEGE OF ENGINEERING AND TECHNOLOGY. A seminar report on GENETIC ALGORITHMS. VISHVESHWARAIAH TECHNOLOGICAL UNIVERSITY S.D.M COLLEGE OF ENGINEERING AND TECHNOLOGY A seminar report on GENETIC ALGORITHMS Submitted by Pranesh S S 2SD06CS061 8 th semester DEPARTMENT OF COMPUTER SCIENCE

More information

Minimizing Makespan for Machine Scheduling and Worker Assignment Problem in Identical Parallel Machine Models Using GA

Minimizing Makespan for Machine Scheduling and Worker Assignment Problem in Identical Parallel Machine Models Using GA , June 30 - July 2, 2010, London, U.K. Minimizing Makespan for Machine Scheduling and Worker Assignment Problem in Identical Parallel Machine Models Using GA Imran Ali Chaudhry, Sultan Mahmood and Riaz

More information

Comparative Study of Different Selection Techniques in Genetic Algorithm

Comparative Study of Different Selection Techniques in Genetic Algorithm Journal Homepage: Comparative Study of Different Selection Techniques in Genetic Algorithm Saneh Lata Yadav 1 Asha Sohal 2 Keywords: Genetic Algorithms Selection Techniques Roulette Wheel Selection Tournament

More information

Chapter 13 Outline. Challenge: Intel and AMD s Advertising Strategies. An Overview of Game Theory. An Overview of Game Theory

Chapter 13 Outline. Challenge: Intel and AMD s Advertising Strategies. An Overview of Game Theory. An Overview of Game Theory Chapter 13 Game Theory A camper awakens to the growl of a hungry bear and sees his friend putting on a pair of running shoes. You can t outrun a bear, scoffs the camper. His friend coolly replies, I don

More information

College of information technology Department of software

College of information technology Department of software University of Babylon Undergraduate: third class College of information technology Department of software Subj.: Application of AI lecture notes/2011-2012 ***************************************************************************

More information

Genetic Programming for Symbolic Regression

Genetic Programming for Symbolic Regression Genetic Programming for Symbolic Regression Chi Zhang Department of Electrical Engineering and Computer Science, University of Tennessee, Knoxville, TN 37996, USA Email: czhang24@utk.edu Abstract Genetic

More information

Behavioral Biases in Auctions: an Experimental Study $

Behavioral Biases in Auctions: an Experimental Study $ Behavioral Biases in Auctions: an Experimental Study $ Anna Dodonova School of Management, University of Ottawa, 136 Jean-Jacques Lussier, Ottawa, ON, K1N 6N5, Canada. Tel.: 1-613-562-5800 ext.4912. Fax:

More information

Logistics. Final exam date. Project Presentation. Plan for this week. Evolutionary Algorithms. Crossover and Mutation

Logistics. Final exam date. Project Presentation. Plan for this week. Evolutionary Algorithms. Crossover and Mutation Logistics Crossover and Mutation Assignments Checkpoint -- Problem Graded -- comments on mycourses Checkpoint --Framework Mostly all graded -- comments on mycourses Checkpoint -- Genotype / Phenotype Due

More information

Mechanism Design in Social Networks

Mechanism Design in Social Networks Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence (AAAI-17) Mechanism Design in Social Networks Bin Li, a Dong Hao, a Dengji Zhao, b Tao Zhou a a Big Data Research Center, University

More information

Outline. Protocols, Strategies and Architectures for Automated Negotiation. Claudio Bartolini HP Labs Bristol Bologna, November 17, 2000

Outline. Protocols, Strategies and Architectures for Automated Negotiation. Claudio Bartolini HP Labs Bristol Bologna, November 17, 2000 Protocols, Strategies and Architectures for Automated Negotiation Claudio Bartolini HP Labs Bristol Bologna, November 17, 2000 Page 1 Outline Part I Definition of the Negotiation Process Part II Importance

More information

Autonomous Agents and Multi-Agent Systems* 2015/2016. Lecture Reaching Agreements

Autonomous Agents and Multi-Agent Systems* 2015/2016. Lecture Reaching Agreements Autonomous Agents and Multi-Agent Systems* 2015/2016 Lecture Reaching Agreements Manuel LOPES * These slides are based on the book by Prof. M. Wooldridge An Introduction to Multiagent Systems and the online

More information

ATTac-2000: An Adaptive Autonomous Bidding Agent. Greg Miller, Cody Musilek, John Steinbach

ATTac-2000: An Adaptive Autonomous Bidding Agent. Greg Miller, Cody Musilek, John Steinbach ATTac-2000: An Adaptive Autonomous Bidding Agent Greg Miller, Cody Musilek, John Steinbach Source Paper Stone, P., M. L. Littman, S. Singh, and M. Kearns (2001). ATTac-2000: An Adaptive Autonomous Bidding

More information

Storage Unit Auctions Buy Bins Like the Pros!

Storage Unit Auctions Buy Bins Like the Pros! Storage Unit Auctions Buy Bins Like the Pros! This e-book is for informational purposes only. It, in no way, guarantees success in buying bins from storage unit auctions or reselling an auctioned bin s

More information

Recap Beyond IPV Multiunit auctions Combinatorial Auctions Bidding Languages. Multi-Good Auctions. CPSC 532A Lecture 23.

Recap Beyond IPV Multiunit auctions Combinatorial Auctions Bidding Languages. Multi-Good Auctions. CPSC 532A Lecture 23. Multi-Good Auctions CPSC 532A Lecture 23 November 30, 2006 Multi-Good Auctions CPSC 532A Lecture 23, Slide 1 Lecture Overview 1 Recap 2 Beyond IPV 3 Multiunit auctions 4 Combinatorial Auctions 5 Bidding

More information

Towards An Automated Multiagent Negotiation System Based On FIPA Specifications

Towards An Automated Multiagent Negotiation System Based On FIPA Specifications 6th WSEAS International Conference on CIRCUITS, SYSTEMS, ELECTRONICS,CONTROL & SIGNAL PROCESSING, Cairo, Egypt, Dec 29-31, 2007 603 Towards An Automated Multiagent Negotiation System Based On FIPA Specifications

More information

Buy-It-Now or Snipe on ebay?

Buy-It-Now or Snipe on ebay? Association for Information Systems AIS Electronic Library (AISeL) ICIS 2003 Proceedings International Conference on Information Systems (ICIS) December 2003 Buy-It-Now or Snipe on ebay? Ilke Onur Kerem

More information

MARKETS AND SOCIETY Microeconomics in Context (Goodwin, et al.), 3 rd Edition

MARKETS AND SOCIETY Microeconomics in Context (Goodwin, et al.), 3 rd Edition Chapter 2 MARKETS AND SOCIETY Microeconomics in Context (Goodwin, et al.), 3 rd Edition Chapter Overview This chapter points out that the economy may be understood as existing within three spheres of activity;

More information

Introduction to Auctions

Introduction to Auctions Common to Optimal in Item Cheriton School of Computer Science University of Waterloo Outline Common Optimal in Item 1 2 Common Optimal 3 in 4 Item 5 Common Optimal in Item Methods for allocating goods,

More information

TRANSPORTATION PROBLEM AND VARIANTS

TRANSPORTATION PROBLEM AND VARIANTS TRANSPORTATION PROBLEM AND VARIANTS Introduction to Lecture T: Welcome to the next exercise. I hope you enjoyed the previous exercise. S: Sure I did. It is good to learn new concepts. I am beginning to

More information

Federal Communications Commission ComCom Federal Office of Communications Worked Example - CCA

Federal Communications Commission ComCom Federal Office of Communications Worked Example - CCA Federal Communications Commission ComCom Federal Office of Communications Worked Example - CCA Addendum to the invitation to tender for frequency blocks for the national provision of mobile telecommunication

More information

Topic 2: Accounting Information for Decision Making and Control

Topic 2: Accounting Information for Decision Making and Control Learning Objectives BUS 211 Fall 2014 Topic 1: Introduction Not applicable Topic 2: Accounting Information for Decision Making and Control State and describe each of the 4 items in the planning and control

More information

Intelligent Techniques Lesson 4 (Examples about Genetic Algorithm)

Intelligent Techniques Lesson 4 (Examples about Genetic Algorithm) Intelligent Techniques Lesson 4 (Examples about Genetic Algorithm) Numerical Example A simple example will help us to understand how a GA works. Let us find the maximum value of the function (15x - x 2

More information

An Agent Model for First Price and Second Price Private Value Auctions

An Agent Model for First Price and Second Price Private Value Auctions An Agent Model for First Price and Second Price Private Value Auctions A. J. Bagnall and I. Toft School of Information Systems University of East Anglia Norwich, NR47TJ England ajb@sys.uea.ac.uk Abstract.

More information

Intro to Algorithmic Economics, Fall 2013 Lecture 1

Intro to Algorithmic Economics, Fall 2013 Lecture 1 Intro to Algorithmic Economics, Fall 2013 Lecture 1 Katrina Ligett Caltech September 30 How should we sell my old cell phone? What goals might we have? Katrina Ligett, Caltech Lecture 1 2 How should we

More information

2. Genetic Algorithms - An Overview

2. Genetic Algorithms - An Overview 2. Genetic Algorithms - An Overview 2.1 GA Terminology Genetic Algorithms (GAs), which are adaptive methods used to solve search and optimization problems, are based on the genetic processes of biological

More information

The Efficient Allocation of Individuals to Positions

The Efficient Allocation of Individuals to Positions The Efficient Allocation of Individuals to Positions by Aanund Hylland and Richard Zeckhauser Presented by Debreu Team: Justina Adamanti, Liz Malm, Yuqing Hu, Krish Ray Hylland and Zeckhauser consider

More information

Predictive Planning for Supply Chain Management

Predictive Planning for Supply Chain Management Predictive Planning for Supply Chain Management David Pardoe and Peter Stone Department of Computer Sciences The University of Texas at Austin {dpardoe, pstone}@cs.utexas.edu Abstract Supply chains are

More information

Final Exam Solutions

Final Exam Solutions 14.27 Economics and E-Commerce Fall 14 Final Exam Solutions Prof. Sara Ellison MIT OpenCourseWare 1. a) Some examples include making prices hard to find, offering several versions of a product differentiated

More information

BUILDING BUSINESS HEURISTICS WITH DATA-MINING INTERNET AGENTS

BUILDING BUSINESS HEURISTICS WITH DATA-MINING INTERNET AGENTS BUILDING BUSINESS HEURISTICS WITH DATA-MINING INTERNET AGENTS Steven Walczak University of Colorado, Denver swalczak@carbon.cudenver.edu Dawn G. Gregg University of Coloraod, Denver dawn.gregg@cudenver.edu

More information

Interleaving Cryptography and Mechanism Design: the Case of Online Auctions. Edith Elkind Helger Lipmaa

Interleaving Cryptography and Mechanism Design: the Case of Online Auctions. Edith Elkind Helger Lipmaa Interleaving Cryptography and Mechanism Design: the Case of Online Auctions Edith Elkind Helger Lipmaa Auctions: Typical Setting One seller, one object, n buyers. Buyers value the object differently private

More information

Introduction to Genetic Algorithm (GA) Presented By: Rabiya Khalid Department of Computer Science

Introduction to Genetic Algorithm (GA) Presented By: Rabiya Khalid Department of Computer Science Introduction to Genetic Algorithm (GA) Presented By: Rabiya Khalid Department of Computer Science 1 GA (1/31) Introduction Based on Darwin s theory of evolution Rapidly growing area of artificial intelligence

More information

An Application of E-Commerce in Auction Process

An Application of E-Commerce in Auction Process An Application of E-Commerce in Auction Process MARIO SPUNDAK, VEDRAN BATOS, MARIO MILICEVIC Department of Electrical Engineering and Computing University of Dubrovnik Cira Carica 4, Dubrovnik 20000 CROATIA

More information

Detecting and Pruning Introns for Faster Decision Tree Evolution

Detecting and Pruning Introns for Faster Decision Tree Evolution Detecting and Pruning Introns for Faster Decision Tree Evolution Jeroen Eggermont and Joost N. Kok and Walter A. Kosters Leiden Institute of Advanced Computer Science Universiteit Leiden P.O. Box 9512,

More information

Evolutionary Algorithms

Evolutionary Algorithms Evolutionary Algorithms Fall 2008 1 Introduction Evolutionary algorithms (or EAs) are tools for solving complex problems. They were originally developed for engineering and chemistry problems. Much of

More information

The Sealed Bid Auction Experiment:

The Sealed Bid Auction Experiment: The Sealed Bid Auction Experiment: This is an experiment in the economics of decision making. The instructions are simple, and if you follow them carefully and make good decisions, you may earn a considerable

More information

Miscomputing Ratio: The Social Cost of Selfish Computing

Miscomputing Ratio: The Social Cost of Selfish Computing Miscomputing Ratio: The Social Cost of Selfish Computing Kate Larson and Tuomas Sandholm Computer Science Department Carnegie Mellon University 5000 Forbes Ave Pittsburgh, PA 15213 {klarson,sandholm}@cs.cmu.edu

More information

A Genetic Algorithm on Inventory Routing Problem

A Genetic Algorithm on Inventory Routing Problem A Genetic Algorithm on Inventory Routing Problem Artvin Çoruh University e-mail: nevin.aydin@gmail.com Volume 3 No 3 (2014) ISSN 2158-8708 (online) DOI 10.5195/emaj.2014.31 http://emaj.pitt.edu Abstract

More information

Market-Based Transmission Expansion Planning

Market-Based Transmission Expansion Planning Energy and Power Engineering, 2012, 4, 387-391 http://dx.doi.org/10.4236/epe.2012.46051 Published Online November 2012 (http://www.scirp.org/journal/epe) Market-Based Transmission Expansion Planning Pavel

More information

A game is a collection of players, the actions those players can take, and their preferences over the selection of actions taken by all the players

A game is a collection of players, the actions those players can take, and their preferences over the selection of actions taken by all the players Game theory review A game is a collection of players, the actions those players can take, and their preferences over the selection of actions taken by all the players A strategy s i is dominant for player

More information

Software Next Release Planning Approach through Exact Optimization

Software Next Release Planning Approach through Exact Optimization Software Next Release Planning Approach through Optimization Fabrício G. Freitas, Daniel P. Coutinho, Jerffeson T. Souza Optimization in Software Engineering Group (GOES) Natural and Intelligent Computation

More information

An Evaluation of Communication Demand of Auction Protocols in Grid Environments

An Evaluation of Communication Demand of Auction Protocols in Grid Environments An Evaluation of Communication Demand of Auction Protocols in Grid Environments MARCOS DIAS DE ASSUNÇÃO AND RAJKUMAR BUYYA Grid Computing and Distributed Systems Laboratory and NICTA Victoria Laboratory

More information

Introduction To Genetic Algorithms

Introduction To Genetic Algorithms 1 Introduction To Genetic Algorithms Dr. Rajib Kumar Bhattacharjya Department of Civil Engineering IIT Guwahati Email: rkbc@iitg.ernet.in References 2 D. E. Goldberg, Genetic Algorithm In Search, Optimization

More information

Robust Multi-unit Auction Protocol against False-name Bids

Robust Multi-unit Auction Protocol against False-name Bids 17th International Joint Conference on Artificial Intelligence (IJCAI-2001) Robust Multi-unit Auction Protocol against False-name Bids Makoto Yokoo, Yuko Sakurai, and Shigeo Matsubara NTT Communication

More information

CHAPTER 5 SOCIAL WELFARE MAXIMIZATION FOR HYBRID MARKET

CHAPTER 5 SOCIAL WELFARE MAXIMIZATION FOR HYBRID MARKET 61 CHAPTER 5 SOCIAL WELFARE MAXIMIZATION FOR HYBRID MARKET 5.1 INTRODUCTION Electricity markets throughout the world continue to be opened to competitive forces. The underlying objective of introducing

More information

Multiagent Systems: Spring 2006

Multiagent Systems: Spring 2006 Multiagent Systems: Spring 2006 Ulle Endriss Institute for Logic, Language and Computation University of Amsterdam Ulle Endriss (ulle@illc.uva.nl) 1 Combinatorial Auctions In a combinatorial auction, the

More information

MAXIMIZE PROFITS VAR BUSINESS

MAXIMIZE PROFITS VAR BUSINESS HOW TO MAXIMIZE PROFITS IN A LOW MARGIN VAR BUSINESS Introduction With each passing day, VAR business is getting more and more competitive. As a result, margins have shrunk considerably which has reduced

More information

Introduction to Multi-Agent Programming

Introduction to Multi-Agent Programming Introduction to Multi-Agent Programming 11. Auctions English, Dutch, Vickrey, and Combinatorial Auctions Alexander Kleiner, Bernhard Nebel Contents Introduction Auction Parameters English, Dutch, and Vickrey

More information

Genetic Algorithm for Predicting Protein Folding in the 2D HP Model

Genetic Algorithm for Predicting Protein Folding in the 2D HP Model Genetic Algorithm for Predicting Protein Folding in the 2D HP Model A Parameter Tuning Case Study Eyal Halm Leiden Institute of Advanced Computer Science, University of Leiden Niels Bohrweg 1 2333 CA Leiden,

More information

GENETIC ALGORITHM BASED APPROACH FOR THE SELECTION OF PROJECTS IN PUBLIC R&D INSTITUTIONS

GENETIC ALGORITHM BASED APPROACH FOR THE SELECTION OF PROJECTS IN PUBLIC R&D INSTITUTIONS GENETIC ALGORITHM BASED APPROACH FOR THE SELECTION OF PROJECTS IN PUBLIC R&D INSTITUTIONS SANJAY S, PRADEEP S, MANIKANTA V, KUMARA S.S, HARSHA P Department of Human Resource Development CSIR-Central Food

More information

CHAPTER 2. Presented By: Raghda Essam Dina El-Haddad Samar El-Haddad Menna Hatem. E-Marketplaces: Structures, Mechanisms and Impacts

CHAPTER 2. Presented By: Raghda Essam Dina El-Haddad Samar El-Haddad Menna Hatem. E-Marketplaces: Structures, Mechanisms and Impacts CHAPTER 2 Presented By: Raghda Essam Dina El-Haddad Samar El-Haddad Menna Hatem E-Marketplaces: Structures, Mechanisms and Impacts Agenda Introduction E-Marketplaces and Its components Transaction, Electronic

More information

HIERARCHICAL decision making problems arise naturally

HIERARCHICAL decision making problems arise naturally IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING, VOL. 5, NO. 3, JULY 2008 377 Mechanism Design for Single Leader Stackelberg Problems and Application to Procurement Auction Design Dinesh Garg and

More information

Plan for today GENETIC ALGORITHMS. Randomised search. Terminology: The GA cycle. Decoding genotypes

Plan for today GENETIC ALGORITHMS. Randomised search. Terminology: The GA cycle. Decoding genotypes GENETIC ALGORITHMS Jacek Malec email: jacek.malec@cs.lth.se Plan for today What is a genetic algorithm? Degrees of freedom. Some examples. Co-evolution, SAGA, Genetic Programming, Evolutionary Strategies,...

More information

CS 161: E-commerce. Stages in E-commerce purchase. Stages in e-commerce purchase. Credit cards as an enabler. Why is a credit card transaction 50?

CS 161: E-commerce. Stages in E-commerce purchase. Stages in e-commerce purchase. Credit cards as an enabler. Why is a credit card transaction 50? 2005 by J. D. Tygar, cs.161.org, 24 Oct 2005 1 CS 161: E-commerce Stages in E-commerce purchase October 24, 2005 2005 by J. D. Tygar, cs.161.org, 24 Oct 2005 2 Stages in e-commerce purchase Advertising

More information

arxiv: v2 [cs.ai] 15 Apr 2014

arxiv: v2 [cs.ai] 15 Apr 2014 Auction optimization with models learned from data arxiv:1401.1061v2 [cs.ai] 15 Apr 2014 Sicco Verwer Yingqian Zhang Qing Chuan Ye Abstract In a sequential auction with multiple bidding agents, it is highly

More information

Keyword Analysis. Section 1: Google Forecast. Example, inc

Keyword Analysis. Section 1: Google Forecast. Example, inc Section 1: Google Forecast Below is a forecast for some example leather keywords that we recommend focus on. Before the forecast is viewed, keep in mind, that Google bases its numbers on average performance

More information