The Asymmetric Network Equilibrium Problem

Here you can get 7 test networks for the asymmetric network equilibrium problem:

Network	nodes	arcs	O/D pairs	Topology	O/D demands	Arc costs
Betsekas-Gafni	25	40	5	bergaf_net	bergaf_demands	bergaf_costs
Nagurney 1	20	28	8	nag1_net	nag1_demands	nag1_costs
Nagurney 2	22	36	12	nag2_net	nag2_demands	nag2_costs
Nagurney 3	25	37	6	nag3_net	nag3_demands	nag3_costs
Nagurney 4	40	66	6	nag4_net	nag4_demands	nag4_costs
Asymmetric Sioux Falls	24	76	528	sf_net	sf_demands	sf_costs
Asymmetric Barcelona	1,020	2,522	7,922	barcelona_net	barcelona_demands	barcelona_costs
Asymmetric Winnipeg	1,052	2,836	4,345	winnipeg_net	winnipeg_demands	winnipeg_costs
Arezzo	213	598	2,423	arezzo_net	arezzo_demands	arezzo_costs
Lazio	306	926	5,683	lazio_net	lazio_demands	lazio_costs

The networks Bertsekas-Gafni, Nagurney 1, 2, 3, and 4 are well known in literature [1, 2] and here described by 3 files:

topology contains one row per arc: tail node – head node
O/D demands contains one row per O/D pair: origin – destination – demand
arc costs is a MATLAB m-file defining the cost function

The networks Asymmetric Sioux Falls, Asymmetric Bercelona, and Asymmetric Winnipeg are modifications of well known networks in literature [3, 4] in which the separable arc cost function is substituted with the following arc cost function with asymmetric jacobian:

c_i,j = t_i,j*{ 1 + B_i,j*[ (f_i,j + K_i,j*f_j,i) / (2*C_i,j) ]^p_i,j }, (1) where:

c_i,j = travel cost on arc (i,j)
t_i,j = free flow time on arc (i,j)
B_i,j = constant B on arc (i,j)
f_i,j = flow on arc (i,j)
K_i,j = asymmetry factor on arc (i,j)
f_j,i = flow on arc (j,i) opposite of (i,j)
C_i,j = capacity of arc (i,j)
p_i,j = power on arc (i,j)

These networks are described by 3 text files:

topology contains one row per arc: tail node – head node
O/D demands contains one row per O/D pair: origin – destination – demand
arc costs contains one row per arc: free flow time – constant B – capacity – power – asymmetry factor

The networks Arezzo and Lazio are new real networks representing the extraurban area of the city of Arezzo (Italy) and of the region Lazio (Italy). They have been provided by Prof. Antonio Pratelli at the Department of Civil Engineering, University of Pisa.
These networks have arc cost functions defined as in (1) and they are described by 3 text files:

topology contains one row per arc: tail node – head node
O/D demands contains one row per O/D pair: origin – destination – demand
arc costs contains one row per arc: free flow time – constant B – capacity – power – asymmetry factor

References

Bertsekas D. P. and Gafni E. M., Projection methods for variational inequalities with application to the traffic assignment problem, Mathematical Programming Study, vol. 17 (1982), pp. 139?159.
Nagurney A., Comparative test of multimodal traffic equilibrium methods, Transportation Research B, vol. 18 (1984), pp. 469?485.
LeBlanc L. J., Morlok E. K. and Pierskalla W. P., An Efficient Approach to Solving the Road Network Equilibrium Traffic Assignment Problem, Transportation Research, vol. 9 (1975), pp. 309?318.
http://www.bgu.ac.il/~bargera/tntp
Panicucci B., Pappalardo M., Passacantando M. (2007), A path-based double projection method for solving the asymmetric traffic network equilibrium problem, Optimization Letters, vol. 1, pp. 171-185