MPE Mathematical Problems in Engineering 1563-5147 1024-123X Hindawi Publishing Corporation 640960 10.1155/2014/640960 640960 Research Article A Mixed 0-1 Linear Programming Approach to the Computation of All Pure-Strategy Nash Equilibria of a Finite n-Person Game in Normal Form Wu Zhengtian 1, 2 Dang Chuangyin 2 Karimi Hamid Reza 3 Zhu Changan 1 Gao Qing 4 Lütjen Michael 1 Department of Precision Machinery and Precision Instrumentation University of Science and Technology of China Hefei 230026 China ustc.edu 2 Department of Systems Engineering and Engineering Management City University of Hong Kong Hong Kong cityu.edu.hk 3 Department of Engineering, Faculty of Engineering and Science University of Agder 4898 Grimstad Norway uia.no 4 School of Engineering and Information Technology University of New South Wales at the Australian Defence Force Academy Canberra ACT 2600 Australia unsw.edu.au 2014 842014 2014 19 01 2014 03 03 2014 8 4 2014 2014 Copyright © 2014 Zhengtian Wu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A main concern in applications of game theory is how to effectively select a Nash equilibrium, especially a pure-strategy Nash equilibrium for a finite n -person game in normal form. This selection process often requires the computation of all Nash equilibria. It is well known that determining whether a finite game has a pure-strategy Nash equilibrium is an NP-hard problem and it is difficult to solve by naive enumeration algorithms. By exploiting the properties of pure strategy and multilinear terms in the payoff functions, this paper formulates a new mixed 0-1 linear program for computing all pure-strategy Nash equilibria. To our knowledge, it is the first method to formulate a mixed 0-1 linear programming for pure-strategy Nash equilibria and it may work well for similar problems. Numerical results show that the approach is effective and this method can be easily distributed in a distributed way.

1. Introduction

Nash equilibrium, as defined in , plays a fundamental role in the development of game theory and its diverse applications in areas such as biology, computer science, economics, and management and social sciences. In these applications, a main concern is how to effectively select a Nash equilibrium, especially a pure-strategy Nash equilibrium for a finite n -person game in normal form. To tackle this problem, many contributions have been made in the literature.

In recent years, most related work is focused on finding mixed-strategy Nash equilibrium. In terms of computational complexity, it was shown in [1, 2] that the problem of finding one mixed-strategy Nash equilibrium is PPAD-complete. The first method for computing Nash equilibrium of a two-person game was developed in  and was extended to computing Nash equilibria of n -person games independently in [4, 5]. The existence of Nash equilibrium was established from an application of Brouwer and Kakutani fixed-point theorems in [6, 7]. To approximate fixed points of continuous mappings, simplicial methods were originated in  and substantially developed in the literature such as . Simplicial methods were adapted for computing Nash equilibria in . Several more advanced simplicial methods for computing Nash equilibria were presented in . A pivoting procedure was proposed in [20, 21] for computing the Nash equilibria of two-person games selected by the linear tracing procedure proposed in . This procedure was extended to n -person games in . A survey of the literature on the computation of Nash equilibria before 1996 can be found in  and on computing equilibria for two-person games in . By exploiting differentiability of the problem, several methods have been developed for computing Nash equilibria in the literature. A smooth logarithmic tracing procedure with two parameters was proposed in . It was proved in  that the procedure converges to a Nash equilibrium as two parameters approach zero one after another. The linear tracing procedure together with well-chosen nonlinear transformations of variables of the system of the optimality conditions leads to a differentiable homotopy method in . As a result of Kohlberg and Mertens’ structure theorem , the global Newton method in  was combined with the homotopy method in  to obtain an algorithm for computing Nash equilibrium. A survey on homotopy methods for computing Nash equilibrium is referred to in . Some new related work has been developed in the literature .

All the algorithms mentioned above can only find a sample Nash equilibrium. Unfortunately, in practice, a game can have many Nash equilibria and which equilibrium should be effectively selected to play is a very challenging issue. The equilibrium selection process often requires the computation of all Nash equilibria of a game. To address this issue, several algorithms have been developed in the literature. A path-following algorithm was described in  to compute all Nash equilibria of a bimatrix game. A homotopy algorithm was presented in  to find all Nash equilibria of an n -person game. Both of these algorithms are based on systems of polynomial equations. A survey on polyhedral homotopy continuation methods for all Nash equilibria can be found in . In , a mixed-integer program (MIP) was formulated to find Nash equilibria of a bimatrix game, which can be employed to compute all Nash equilibria of a bimatrix game. Recently, by enumerating vertices of polyhedrons, two algorithms were introduced in  for computing all Nash equilibria of a bimatrix game.

As a special example of the mixed-strategy Nash equilibrium, the pure-strategy Nash equilibrium has gotten more attention recently. The work of  shows that determining whether a game has a pure-strategy Nash equilibrium is NP-hard. The work of  presents a novel generic mapping between graphical games and Markov random fields so that pure Nash equilibria in the former can be found by statistical inference on the latter. Based on a new description format of game-stimulate-response pair, the work of  puts forward a constraints’ satisfaction-based algorithm on this data structure to compute pure Nash equilibrium of graphical game. The work of  develops an efficient algorithm, which is called Valued Nash Propagation, for computing pure strategy Nash equilibria that satisfy various criteria (such as the utilitarian or Nash-Bernoulli social welfare functions) in games with sparse interaction structure. Some other related work can be found in [2, 45, 46].

As a special case of integer programming, the mixed 0-1 linear programming, in which unknowns are binary, has many applications in operation research. Many classic problems, such as the knapsack problem, facility location problem, production planning problem, and time tabling problem, can be formulated as a mixed 0-1 linear programming. The research of these problems which has a long history and many commercial softwares, most of which are based on branch and bound method, cutting plane method, or branch and cut method, could solve this mixed 0-1 linear programming well.

This paper studies the computation of all pure-strategy Nash equilibria of a finite n -person game in normal form. To solve this problem, we formulate a mixed 0-1 linear programming by exploiting the properties of pure strategy and multilinear forms in the payoff functions. An enumeration of all the feasible solutions of the mixed 0-1 linear program yields all pure-strategy Nash equilibria. Numerical results show that the method is promising. To our knowledge, it is the first method to formulate a mixed 0-1 linear programming for pure-strategy Nash equilibria and this method may work well for other similar problems. Furthermore, as an advantage of the formulation, one can apply distributed computation to dramatically speed up the computation.

The rest of the paper is organized as follows. Section 2 formulates a mixed 0-1 linear program for computing all pure-strategy equilibria. Section 3 presents numerical results to show that the approach is effective. The paper is concluded with some remarks in Section 4.

2. A Mixed 0-1 Linear Programming Formulation

In this section a mixed 0-1 linear programming will be formulated to all pure-strategy Nash equilibria of a finite game.

Let N = { 1,2 , , n } be the set of players. The pure strategy set of player i N is denoted by S i = { s j i j M i } with M i = { 1,2 , , m i } . Given S i with i N , the set of all pure strategy profiles is S = i = 1 n S i . We denote the payoff function of player i N by u i : S R . For i N , let S - i = k N { i } S k . Then, s = ( s j 1 1 , s j 2 2 , , s j n n ) S can be rewritten as s = ( s j i i , s - i ) with s - i = ( s j 1 1 , , s j i - 1 i - 1 , s j i + 1 i + 1 , , s j n n ) S - i . A mixed strategy of player i is a probability distribution on S i denoted by x i = ( x 1 i , x 2 i , , x m i i ) . Let X i be the set of all mixed strategies of player i . Then, X i = { x i = ( x 1 i , x 2 i , , x m i i ) R + m i j = 1 m i x j i = 1 } . Thus, for x i X i , the probability assigned to pure strategy s j i S i is equal to x j i . Given X i with i N , the set of all mixed strategy profiles is X = i = 1 n X i . For i N , let X - i = k N { i } X k . Then, x = ( x 1 , x 2 , , x n ) X can be rewritten as x = ( x i , x - i ) with x - i = ( x 1 , , x i - 1 , x i + 1 , , x n ) X - i . If x X is played, then the probability that a pure strategy profile s = ( s j 1 1 , s j 2 2 , , s j n n ) S occurs is i = 1 n x j i i . Therefore, for x X , the expected payoff of player i is given by u i ( x ) = s S u i ( s ) i = 1 n x j i i . With these notations, a finite n -person game in normal form can be represented as Γ = N , S , { u i } i N or Γ = N , X , { u i } i N .

Definition 1 (Nash, [<xref ref-type="bibr" rid="B1">7</xref>]).

A mixed strategy profile x * X is a Nash equilibrium of game Γ if u i ( x * ) u i ( x i , x * - i ) for all i N and x i X i .

With this definition, an application of the optimality condition leads to the fact that x * is a Nash equilibrium if and only if there are λ * and μ * together with x * satisfying the system of (1) u i ( s j i , x - i ) + λ j i - μ i = 0 , e i x i - 1 = 0 , x j i λ j i = 0 , x j i 0 , λ j i 0 , j = 1,2 , , m i , i = 1,2 , , n , where e i = ( 1,1 , , 1 ) R m i .

Lemma 2.

Let β be a given positive number such that (2) β max i N { max s S u i ( s ) - min s S u i ( s ) } . Then, (1) is equivalent to (3) u i ( s j i , x - i ) + λ j i - μ i = 0 , e i x i - 1 = 0 , x j i v j i , λ j i β ( 1 - v j i ) , v j i { 0,1 } , x j i 0 , λ j i 0 , j = 1,2 , , m i , i = 1,2 , , n .

Proof.

Let ( x , λ , v ) be a solution of the system (3). Suppose that v j i = 0 in (3). Then, x j i = 0 and λ j i β . Thus, x j i λ j i = 0 . Suppose that v j i = 1 in (3). Then, λ j i = 0 . Thus, x j i λ j i = 0 . Therefore, ( x , λ ) is a solution of the system (1).

Let ( x , λ ) be a solution of the system (1). Since j = 1 m i x j i = 1 , there must exist some k M i such that x k i > 0 . From x k i λ k i = 0 , we get that λ k i = 0 . Thus, μ i = u i ( s k i , x - i ) max s S u i ( s ) . Therefore, for any j M i , we obtain from the system (1) that (4) λ j i = μ i - u i ( s j i , x - i ) max s S u i ( s ) - min s S u i ( s ) β .

Suppose that x j i = 0 . Let v j i = 0 . Thus, (5) x j i = 0 v j i , λ j i β ( 1 - v j i ) = β . Suppose that x j i > 0 . Let v j i = 1 . Thus, (6) x j i v j i = 1 , λ j i = 0 β ( 1 - v j i ) = 0 . Therefore, ( x , λ , v ) is a solution of the system (3). This completes the proof.

This lemma implies that finding a pure-strategy Nash equilibrium is equivalent to finding a solution to the system of (7) u i ( s j i , x - i ) + λ j i - μ i = 0 , e i x i - 1 = 0 , λ j i β ( 1 - x j i ) , x j i { 0,1 } , λ j i 0 , j = 1,2 , , m i , i = 1,2 , , n . For any s j i S i , one can obtain from u i ( x ) that (8) u i ( s j i , x - i ) = s - i = ( s j 1 1 , , s j i - 1 i - 1 , s j i + 1 i + 1 , , s j n n ) S - i u i ( s j i , s - i ) k i x j k k . Let y ( s - i ) = k i x j k k for s - i = ( s j 1 1 , , s j i - 1 i - 1 , s j i + 1 i + 1 , , s j n n ) S - i . Then, (9) u i ( s j i , x - i ) = s - i = ( s j 1 1 , , s j i - 1 i - 1 , s j i + 1 i + 1 , , s j n n ) S - i u i ( s j i , s - i ) y ( s - i ) . Substituting this into (7) yields (10) s - i S - i u i ( s j i , s - i ) y ( s - i ) + λ j i - μ i = 0 , e i x i - 1 = 0 , λ j i β ( 1 - x j i ) , x j i { 0,1 } , λ j i 0 , j = 1,2 , , m i , i = 1,2 , , n , y ( s - i ) = k i x j k k , s - i S - i , i = 1,2 , , n . Observe that y ( s - i ) { 0,1 } since x j i { 0,1 } . This property leads to the following result.

Lemma 3.

For i N and s - i = ( s j 1 1 , , s j i - 1 i - 1 , s j i + 1 i + 1 , , s j n n ) S - i , the system of (11) y ( s - i ) = k i x j k k , x j k k { 0,1 } , k = 1,2 , , n , k i , is equivalent to the system of (12) y ( s - i ) h i x j h h - ( n - 2 ) , y ( s - i ) x j k k , 0 y ( s - i ) , x j k k { 0,1 } , k = 1,2 , , n , k i .

Proof.

Let ( y , x ) be a solution of the system (12). Suppose that x j k k = 1 , for all k i . Then, (13) y ( s - i ) h i x j h h - ( n - 2 ) = ( n - 1 ) - ( n - 2 ) = 1 , y ( s - i ) x j k k = 1 , k i . Thus, y ( s - i ) = k i x j k k = 1 .

Suppose that x j k k = 0 , k i . Then, (14) y ( s - i ) x j k k = 0 , k i , 0 y ( s - i ) . Thus, y ( s - i ) = k i x j k k = 0 . Therefore, ( y , x ) is a solution of the system (11).

Let ( y , x ) be a solution of the system (11). Suppose that x j k k = 1 , for all k i . Then, y ( s - i ) = k i x j k k = 1 . Thus, (15) h i x j h h - ( n - 2 ) = ( n - 1 ) - ( n - 2 ) = 1 y ( s - i ) = 1 , y ( s - i ) = 1 x j k k = 1 , k i , 0 y ( s - i ) . Suppose that x j k k = 0 , k i . Then, y ( s - i ) = k i x j k k = 0 . Thus, (16) h i x j h h - ( n - 2 ) ( n - 2 ) - ( n - 2 ) = y ( s - i ) = 0 , y ( s - i ) = 0 x j k k , k i , 0 y ( s - i ) . Therefore, ( y , x ) is a solution of the system (12). This completes the proof.

Replacing y ( s - i ) = k i x j k k of the system (10) with the system (12), we obtain a mixed 0-1 linear program, which is as follows: (17) s - i S - i u i ( s j i , s - i ) y ( s - i ) + λ j i - μ i = 0 , e i x i - 1 = 0 , λ j i β ( 1 - x j i ) , x j i { 0,1 } , λ j i 0 , j = 1,2 , , m i , i = 1,2 , , n , y ( s - i ) h i x j h h - ( n - 2 ) , y ( s - i ) x j k k , k i , 0 y ( s - i ) , s - i S - i , i = 1,2 , , n . As a corollary of Lemma 3, we obtain the following result.

Corollary 4.

The system (17) is equivalent to the system (7).

The above idea is illustrated in the following example.

Example 5.

Consider a three-player game Γ = ( N , S , { u i } i N ) , where N = { 1,2 , 3 } , S i = { s 1 i , s 2 i } , i N , and { u i } i N are given by (18) s 1 2 s 2 2 s 1 2 s 2 2 s 1 1 ( 1,1 , 1 ) ( 1,0 , 1 ) ( 0,1 , 0 ) ( 1,0 , 0 ) s 2 1 ( 1,1 , 1 ) ( 1,0 , 1 ) ( 0,1 , 0 ) ( 0,0 , 0 ) s 1 3 s 2 3

Let β = 1 . The mixed 0-1 linear program for this game is given by (19) y ( s 1 2 , s 1 3 ) + y ( s 2 2 , s 1 3 ) + y ( s 2 2 , s 2 3 ) + λ 1 1 - μ 1 = 0 , y ( s 1 2 , s 1 3 ) + y ( s 2 2 , s 1 3 ) + λ 2 1 - μ 1 = 0 , x 1 1 + x 2 1 = 1 , y ( s 1 1 , s 1 3 ) + y ( s 2 1 , s 1 3 ) + y ( s 1 1 , s 2 3 ) + y ( s 2 1 , s 2 3 ) + λ 1 2 - μ 2 = 0 , λ 2 2 - μ 2 = 0 , x 1 2 + x 2 2 = 1 , y ( s 1 1 , s 1 2 ) + y ( s 2 1 , s 1 2 ) + y ( s 1 1 , s 2 2 ) + y ( s 2 1 , s 2 2 ) + λ 1 3 - μ 3 = 0 , λ 2 3 - μ 3 = 0 , x 1 3 + x 2 3 = 1 , λ j i 1 - x j i , x j i { 0,1 } , λ j i 0 , j = 1,2 , i = 1,2 , 3 , y ( s 1 2 , s 1 3 ) x 1 2 + x 1 3 - 1 , y ( s 1 2 , s 1 3 ) x 1 2 , y ( s 1 2 , s 1 3 ) x 1 3 , y ( s 1 2 , s 1 3 ) 0 , y ( s 1 2 , s 2 3 ) x 1 2 + x 2 3 - 1 , y ( s 1 2 , s 2 3 ) x 1 2 , y ( s 1 2 , s 2 3 ) x 2 3 , y ( s 1 2 , s 2 3 ) 0 , y ( s 2 2 , s 1 3 ) x 2 2 + x 1 3 - 1 , y ( s 2 2 , s 1 3 ) x 2 2 , y ( s 2 2 , s 1 3 ) x 1 3 , y ( s 2 2 , s 1 3 ) 0 , y ( s 2 2 , s 2 3 ) x 2 2 + x 2 3 - 1 , y ( s 2 2 , s 2 3 ) x 2 2 , y ( s 2 2 , s 2 3 ) x 2 3 , y ( s 2 2 , s 2 3 ) 0 , y ( s 1 1 , s 1 3 ) x 1 1 + x 1 3 - 1 , y ( s 1 1 , s 1 3 ) x 1 1 , y ( s 1 1 , s 1 3 ) x 1 3 , y ( s 1 1 , s 1 3 ) 0 , y ( s 2 1 , s 1 3 ) x 2 1 + x 1 3 - 1 , y ( s 2 1 , s 1 3 ) x 2 1 , y ( s 2 1 , s 1 3 ) x 1 3 , y ( s 2 1 , s 1 3 ) 0 , y ( s 1 1 , s 2 3 ) x 1 1 + x 2 3 - 1 , y ( s 1 1 , s 2 3 ) x 1 1 , y ( s 1 1 , s 2 3 ) x 2 3 , y ( s 1 1 , s 2 3 ) 0 , y ( s 2 1 , s 2 3 ) x 2 1 + x 2 3 - 1 , y ( s 2 1 , s 2 3 ) x 2 1 , y ( s 2 1 , s 2 3 ) x 2 3 , y ( s 2 1 , s 2 3 ) 0 , y ( s 1 1 , s 1 2 ) x 1 1 + x 1 2 - 1 , y ( s 1 1 , s 1 2 ) x 1 1 , y ( s 1 1 , s 1 2 ) x 1 2 , y ( s 1 1 , s 1 2 ) 0 , y ( s 1 1 , s 2 2 ) x 1 1 + x 2 2 - 1 , y ( s 1 1 , s 2 2 ) x 1 1 , y ( s 1 1 , s 2 2 ) x 2 2 , y ( s 1 1 , s 2 2 ) 0 , y ( s 2 1 , s 1 2 ) x 2 1 + x 1 2 - 1 , y ( s 2 1 , s 1 2 ) x 2 1 , y ( s 2 1 , s 1 2 ) x 1 2 , y ( s 2 1 , s 1 2 ) 0 , y ( s 2 1 , s 2 2 ) x 2 1 + x 2 2 - 1 , y ( s 2 1 , s 2 2 ) x 2 1 , y ( s 2 1 , s 2 2 ) x 2 2 , y ( s 2 1 , s 2 2 ) 0 .

3. Numerical Results

In this section, some numerical results are presented. Under the software Microsoft Visual Studio 2008 and ILOG CPLEX, we use C++ to call ILOG CPLEX API functions which are based on the branch and cut method to solve the mixed 0-1 linear program (17). The MIP (mixed integer programming) search method, which is a dynamic search strategy or branch and cut strategy in ILOG CPLEX, is determined by the ILOG CPLEX automatically. All other parameters are also set automatically by ILOG CPLEX itself.

We have run our code on a workstation of Lenovo ThinkStation D20 4155-BM4 with 16 processors and 16 G RAM. In the presentation of numerical results, some symbols are explained as follows:

Num N : the number of players;

Num S : the number of strategies for each player;

NumEquilibra: the number of pure-strategy Nash equilibria for the instance;

Time: the total computational time to solve the problem.

Example 1.

Consider a three-player game Γ = ( N , S , { u i } i N ) , where N = { 1,2 , 3 } . The number of strategies for each player, Num S , is generated from 3 to 45 randomly. The { u i } i N are generated randomly too. There are three different ranges for { u i } i N in this example, which are from 0 to 10, from 0 to 50, and from 0 to 100.

Let β = 1000 ; 30 instances have been solved for each range of { u i } i N . Numerical results for each range are presented in Tables 1, 2, and 3, respectively.

The payoff function u i ( s j i , s - i ) is generated from 0 to 10 randomly.

Prob. Num N Num S NumEquilibra Time (s)
1 3 4 1 0.53
2 3 4 2 0.62
3 3 4 2 0.56
4 3 5 3 0.56
5 3 6 2 0.41
6 3 6 1 0.66
7 3 7 5 0.59
8 3 8 6 0.58
9 3 8 3 0.89
10 3 8 3 0.81
11 3 10 2 1.00
12 3 11 3 1.47
13 3 12 4 1.81
14 3 14 4 4.91
15 3 14 8 3.81
16 3 15 6 5.87
17 3 19 13 27.86
18 3 19 3 33.40
19 3 21 13 47.28
20 3 23 10 102.51
21 3 24 22 122.94
22 3 25 21 239.20
23 3 26 13 278.50
24 3 26 15 301.58
25 3 28 21 419.97
26 3 33 22 1870.24
27 3 35 20 2082.34
28 3 40 24 4599.72
29 3 44 20 3033.29
30 3 45 26 16992.10

The payoff function u i ( s j i , s - i ) is generated from 0 to 50 randomly.

Prob. Num N Num S NumEquilibra Time (s)
1 3 3 4 0.46
2 3 3 0 0.50
3 3 3 1 0.40
4 3 4 0 0.34
5 3 4 2 0.52
6 3 6 0 0.45
7 3 7 1 0.70
8 3 7 2 0.82
9 3 7 1 0.57
10 3 8 0 0.73
11 3 9 0 0.66
12 3 10 0 0.93
13 3 11 1 1.64
14 3 16 0 13.26
15 3 19 0 33.22
16 3 20 3 40.67
17 3 21 2 55.10
18 3 22 4 74.20
19 3 24 0 195.78
20 3 24 2 142.14
21 3 27 3 507.44
22 3 29 2 1178.18
23 3 30 1 1287.87
24 3 30 1 1667.70
25 3 32 1 1638.94
26 3 33 2 1761.86
27 3 34 0 3125.08
28 3 35 5 3446.67
29 3 44 3 36601.30
30 3 45 5 43583.30

The payoff function u i ( s j i , s - i ) is generated from 0 to 100 randomly.

Prob. Num N Num S NumEquilibra Time (s)
1 3 3 0 0.41
2 3 3 2 0.45
3 3 4 2 0.51
4 3 5 0 0.49
5 3 5 2 0.61
6 3 7 0 0.56
7 3 7 2 0.41
8 3 8 0 0.61
9 3 8 2 0.89
10 3 8 2 0.68
11 3 10 1 1.11
12 3 12 1 4.00
13 3 13 1 2.64
14 3 13 2 2.64
15 3 16 0 14.25
16 3 19 4 25.70
17 3 22 4 55.94
18 3 24 2 132.87
19 3 24 1 142.96
20 3 25 2 246.76
21 3 26 2 499.40
22 3 28 0 804.76
23 3 28 1 671.67
24 3 30 2 1382.03
25 3 32 1 2821.16
26 3 34 2 3325.67
27 3 37 1 7088.06
28 3 38 2 9395.13
29 3 41 2 20441.10
30 3 43 0 29204.70
Example 2.

Consider a five-player game Γ = ( N , S , { u i } i N ) , where N = { 1,2 , 3,4 , 5 } . The number of strategies for each player, Num S , is generated from 3 to 7 randomly. The { u i } i N are generated randomly too. There are three different ranges for { u i } i N in this example, which are from 0 to 10, from 0 to 50, and from 0 to 100.

Let β = 1000 ; 10 instances have been solved for each range of { u i } i N . Numerical results for each range are presented in Tables 4, 5, and 6, respectively.

The payoff function u i ( s j i , s - i ) is generated from 0 to 10 randomly.

Prob. Num N Num S NumEquilibra Time (s)
1 5 3 4 1.09
2 5 3 3 1.18
3 5 3 2 1.19
4 5 4 3 6.85
5 5 4 4 5.26
6 5 5 2 98.95
7 5 5 3 152.76
8 5 5 1 122.88
9 5 6 4 2092.13
10 5 6 2 1693.61

The payoff function u i ( s j i , s - i ) is generated from 0 to 50 randomly.

Prob. Num N Num S NumEquilibra Time (s)
1 5 3 3 1.20
2 5 3 1 0.99
3 5 3 1 1.71
4 5 4 1 5.69
5 5 4 1 5.62
6 5 4 1 11.90
7 5 4 3 6.77
8 5 5 1 1996.60
9 5 6 2 2031.19
10 5 7 3 18094.80

The payoff function u i ( s j i , s - i ) is generated from 0 to 100 randomly.

Prob. Num N Num S NumEquilibra Time (s)
1 5 3 0 0.80
2 5 4 0 10.98
3 5 4 0 11.11
4 5 4 1 6.93
5 5 4 0 11.07
6 5 5 1 99.30
7 5 5 0 138.33
8 5 6 1 2408.44
9 5 6 2 1960.49
10 5 7 3 22237.60

From Tables 1, 2, and 3, one can see that this mixed 0-1 linear programming method can find all pure-strategy Nash equilibria of three-person game effectively. For example, in Table 1 the problem 30 with 45 strategies can be solved in 16992.10 seconds and the number of equilibria is 26. It has been shown that every n -person game can be reduced to a three-person game in polynomial time in . However, some five-person games are still solved by this method in our examples in order to test the effectiveness of this mixed 0-1 linear programming approach. The numerical results for five-person can be obtained from Tables 4, 5, and 6. All the numerical results show that this new method can handle all pure-strategy Nash equilibria of n -person successfully. However, the computation time increases quickly with the increasing of the problems’ dimension. To tackle this problem, the distributed computation, which is a salient feature of the integer programming, will be considered in our next work. Besides, some similar problems with multilinear terms could be solved by this method too.

4. Conclusions

In this paper, a mixed 0-1 linear programming approach has been proposed to find all pure-strategy Nash equilibria of a finite n -person game in normal form. This method is based on the properties of pure strategy and multilinear terms in the payoff functions. Some numerical results have been given too. The results are promising. However, there is still some interesting future work to be done in the next step. For one thing, as a nice feature of integer programming, the computation will be implemented in a distributed way, which can significantly speed up the efficiency. For another, more research about the properties of strategy and multilinear terms has to be done in order to compute more general Nash equilibria, mixed-strategy Nash equilibria.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgment

This work was partially supported by the RCN (Research Council of Norway) DECOFF project number 225231/O70 and by GRF: CityU 112910 of Hong Kong SAR Government.

Chen X. Deng X. Settling the complexity of two-player Nash equilibrium Proceedings of the 47th Annual IEEE Symposium on Foundations of Computer Science (FOCS '06) October 2006 Berkeley, Calif, USA 261 270 2-s2.0-36448948767 10.1109/FOCS.2006.69 Daskalakis C. Goldberg P. W. Papadimitriou C. H. The complexity of computing a Nash equilibrium Proceedings of the 38th Annual ACM Symposium on Theory of Computing (STOC '06) 2006 New York, NY, USA ACM 71 78 10.1145/1132516.1132527 MR2277132 Lemke C. E. Howson J. T. Equilibrium points of bimatrix games Journal of the Society for Industrial and Applied Mathematics 1964 12 413 423 MR0173556 ZBL0128.14804 Rosenmüller J. On a generalization of the Lemke-Howson algorithm to noncooperative n -person games SIAM Journal on Applied Mathematics 1971 21 73 79 MR0287921 10.1137/0121010 Wilson R. Computing equilibria of n -person games SIAM Journal on Applied Mathematics 1971 21 80 87 MR0371424 10.1137/0121011 Nash J. F. Equilibrium points in n -person games Proceedings of the National Academy of Sciences of the United States of America 1950 36 48 49 MR0031701 10.1073/pnas.36.1.48 Nash J. F. Non-cooperative games Annals of Mathematics 1951 54 2 286 295 MR0043432 10.2307/1969529 ZBL0045.08202 Scarf H. The approximation of fixed points of a continuous mapping SIAM Journal on Applied Mathematics 1967 15 1328 1343 MR0242483 10.1137/0115116 ZBL0153.49401 Allgower E. L. Georg K. Piecewise linear methods for nonlinear equations and optimization Journal of Computational and Applied Mathematics 2000 124 1-2 245 261 10.1016/S0377-0427(00)00427-1 MR1803303 ZBL0970.65055 Dang C. Y. The D1-triangulation of Rn for simplicial algorithms for computing solutions of nonlinear equations Mathematics of Operations Research 1991 16 1 148 161 10.1287/moor.16.1.148 MR1106794 Eaves B. C. Homotopies for computation of fixed points Mathematical Programming 1972 3 1 22 MR0303953 ZBL0276.55004 Garcia C. B. Zangwill W. I. Pathways to Solutions, Fixed Points, and Equilibria 1981 Englewood Cliffs, NJ, USA Prentice Hall Kojima M. Yamamoto Y. A unified approach to the implementation of several restart fixed point algorithms and a new variable dimension algorithm Mathematical Programming 1984 28 3 288 328 10.1007/BF02612336 MR736295 ZBL0539.65036 van der Laan G. Talman A. J. J. A restart algorithm for computing fixed points without an extra dimension Mathematical Programming 1979 17 1 74 84 10.1007/BF01588226 MR538124 ZBL0411.90061 Todd M. J. The Computation of Fixed Points and Applications 1976 1 Berlin, Germany Springer Lecture Notes in Economics and Mathematical Systems MR0410732 Garcia C. B. Lemke C. E. Luethi H. Simplicial approximation of an equilibrium point of noncooperative n -person games Mathematical Programming 1973 4 227 260 Doup T. M. Talman A. J. J. A new simplicial variable dimension algorithm to find equilibria on the product space of unit simplices Mathematical Programming 1987 37 3 319 355 10.1007/BF02591741 MR891129 ZBL0616.90084 Doup T. M. Talman A. J. J. A continuous deformation algorithm on the product space of unit simplices Mathematics of Operations Research 1987 12 3 485 521 10.1287/moor.12.3.485 MR906420 ZBL0653.90078 van der Laan G. Talman A. J. J. van der Heyden L. Simplicial variable dimension algorithms for solving the nonlinear complementarity problem on a product of unit simplices using a general labelling Mathematics of Operations Research 1987 12 3 377 397 10.1287/moor.12.3.377 MR906413 ZBL0638.90096 van den Elzen A. H. Talman A. J. J. A procedure for finding Nash equilibria in bi-matrix games Zeitschrift für Operations Research: Mathematical Methods of Operations Research 1991 35 1 27 43 10.1007/BF01415958 MR1095267 ZBL0729.90093 van den Elzen A. Talman D. An algorithmic approach toward the tracing procedure for bi-matrix games Games and Economic Behavior 1999 28 1 130 145 10.1006/game.1998.0688 MR1699008 ZBL0939.91006 Harsanyi J. C. The tracing procedure: a Bayesian approach to defining a solution for n -person noncooperative games International Journal of Game Theory 1975 4 1-2 61 94 MR0391998 10.1007/BF01766187 Herings P. J.-J. van den Elzen A. Computation of the Nash equilibrium selected by the tracing procedure in n -person games Games and Economic Behavior 2002 38 1 89 117 10.1006/game.2001.0856 MR1875928 McKelvey R. D. McLennan A. Computation of equilibria in finite games Handbook of Computational Economics 1996 1 Amsterdam, The Netherlands North-Holland 87 142 MR1416607 ZBL1126.91304 Von Stengel B. Computing equilibria for two-person games Handbook of Game Theory with Economic Applications 2002 3 1723 1759 2-s2.0-67649370955 10.1016/S1574-0005(02)03008-4 Harsanyi J. C. Selten R. A General Theory of Equilibrium Selection in Games 1988 1 Cambridge, Mass, USA MIT Press MR956053 Schanuel S. H. Simon L. K. Zame W. R. The algebraic geometry of games and the tracing procedure Game Equilibrium Models 1991 2 9 43 Herings P. J.-J. Peeters R. J. A. P. A differentiable homotopy to compute Nash equilibria of n -person games Economic Theory 2001 18 1 159 185 10.1007/PL00004129 MR1839785 Govindan S. Wilson R. A global Newton method to compute Nash equilibria Journal of Economic Theory 2003 110 1 65 86 MR1984254 10.1016/S0022-0531(03)00005-X ZBL1042.91001 Smale S. A convergent process of price adjustment and global Newton methods Journal of Mathematical Economics 1976 3 2 107 120 MR0411577 10.1016/0304-4068(76)90019-7 ZBL0354.90018 Herings P. J.-J. Peeters R. J. A. P. Homotopy methods to compute equilibria in game theory Economic Theory 2010 42 1 119 156 10.1007/s00199-009-0441-5 MR2551731 ZBL1185.91028 Yin S. Ding S. X. Haghani A. Hao H. Zhang P. A comparison study of basic data-driven fault diagnosis and process monitoring methods on the benchmark tennessee eastman process Journal of Process Control 2012 22 9 1567 1581 10.1016/j.jprocont.2012.06.009 Yin S. Ding S. X. Sari A. H. A. Hao H. Data-driven monitoring for stochastic systems and its application on batch process International Journal of Systems Science 2013 44 7 1366 1376 10.1080/00207721.2012.659708 MR3050076 ZBL1278.93259 Zhang H. Shi Y. Wang J. On energy-to-peak filtering for nonuniformly sampled nonlinear systems: a Markovian jump system approach IEEE Transactions on Fuzzy Systems 2014 22 1 212 222 10.1109/TFUZZ.2013.2250291 Yin S. Luo H. Ding S. Real-time implementation of fault-tolerant control systems with performance optimization IEEE Transactions on Industrial Electronics 2013 61 5 2402 2411 Zhang H. Zhang X. Wang J. Robust gain-scheduling energy-to-peak control of vehicle lateral dynamics stabilisation Vehicle System Dynamics 2014 10.1080/00423114.2013.879190 Kostreva M. M. Kinard L. A. A differentiable homotopy approach for solving polynomial optimization problems and noncooperative games Computers & Mathematics with Applications 1991 21 6-7 135 143 10.1016/0898-1221(91)90168-4 MR1096141 ZBL0721.90071 Herings P. J.-J. Peeters R. J. A. P. A globally convergent algorithm to compute all Nash equilibria for n -person games Annals of Operations Research 2005 137 1 349 368 10.1007/s10479-005-2265-4 MR2166447 Datta R. S. Finding all Nash equilibria of a finite game using polynomial algebra Economic Theory 2010 42 1 55 96 10.1007/s00199-009-0447-z MR2551729 ZBL1185.91025 Sandholm T. Gilpin A. Conitzer V. Mixed-integer programming methods for finding Nash equilibria Proceedings of the 20th National Conference on Artificial Intelligence July 2005 495 501 2-s2.0-29344453416 Avis D. Rosenberg G. D. Savani R. von Stengel B. Enumeration of Nash equilibria for two-player games Economic Theory 2010 42 1 9 37 10.1007/s00199-009-0449-x MR2551727 ZBL1182.91013 Gottlob G. Greco G. Scarcello F. Pure Nash equilibria: hard and easy games Proceedings of the 9th Conference on Theoretical Aspects of Rationality and Knowledge 2003 ACM 215 230 Jiang M. Finding pure Nash equilibrium of graphical game via constraints satisfaction approach Combinatorics, Algorithms, Probabilistic and Experimental Methodologies 2007 New York, NY, USA Springer 483 494 Chapman A. C. Farinelli A. de Cote E. M. Rogers A. Jennings N. R. A distributed algorithm for optimising over pure strategy Nash equilibria Proceedings of the 24th AAAI Conference on Artificial Intelligence July 2010 Atlanta, Ga, USA 749 755 2-s2.0-77958568338 Gottlob G. Greco G. Mancini T. Complexity of pure equilibria in bayesian games Proceedings of the International Joint Conference on Artificial Intelligence 2007 1294 1299 Chapman A. C. Control of large distributed systems using games with pure strategy Nash equilibria [Ph.D. thesis] 2009 University of Southampton Bubelis V. On equilibria in finite games International Journal of Game Theory 1979 8 2 65 79 10.1007/BF01768703 MR550952 ZBL0424.90091