|
References | No. of subj. | No. of rounds | No. of routes | Experimental conditionsa | Main contents of analysisa |
|
Iida et al. [33] | 40 | 1: 20, 2: 21 | 2 | Exp. 1: actual TT on the route chosen in the last round Exp. 2: actual and predicted TT on the route chosen in the last round | Traffic flow distribution, actual TT, frequency of route switching, difference between the predicted and actual TT |
Iida et al. [34] | 35 | 63 | 2 | Exp. 1: NI-HQI-HQI Exp. 2: NI-LQI-HQI Exp. 3: NI-LQI-LQI | Route-choice rate in three periods, effect of actual TT information |
Rapoport et al. [28] | 240 | 40 | 2, 3 | Number of subjects between OD increased: 10-20-40 Number of subjects between OD decreased: 40-20-10 | Traffic flow distribution, mean payoff, route switching rate of three routes in experiment B under different OD conditions |
Rapoport et al. [29] | 108 | 1: 40, 2: 80 | 2, 3, 5 | Subjects were divided into 6 groups of 18 members each Three groups participated in condition ADD and three others in condition DELETE. | Traffic flow distribution, mean payoff, number of route switches, frequency of route switching |
Gisches and Rapoport [30] | 180 | 60 | 4, 6 | Subjects were divided into 10 groups of 18 members each. Five groups participated in condition PUBLIC and five others in condition PRIVATE. | Traffic flow distribution, mean payoff, difference of individual route choice, learning behavior |
Mak et al. [32] | 180 | 40, 20 | 2, 3 | Subjects were divided into 10 groups of 18 members each Five groups participated in condition SIM (40 rounds) and five groups in condition SEQ (20 rounds). | Traffic flow distribution under conditions of simultaneous and sequential route choice, subject’s learning behavior |
Selten et al. [35] | 18 | 200 | 2 | Exp. 1: actual TT of the last chosen route Exp. 2: actual TT of the entire road network in the last round | Traffic flow distribution, number of route switches, cumulative payoff |
Avineri and Prashker [36] | 24, 23 | 100 | 2 | Sce. 1 (24 subjects): information acquired by subjects only through their own experience Sce. 2 (23 subjects): a priori static TT information provided to subjects | Average distribution of traffic flow, risk type |
Ben-Elia et al. [37] | 24, 25 | 300 | 2 | Exp. 1 (24 subjects): received real-time and feedback information Exp. 2 (25 subjects): received feedback information Three scenarios in each experiment, including safer-fast, risky-fast, and low-risk. | Statistical analysis of the proportions of choices of the faster route |
Ben-Elia et al. [38] | 36 | 20 | 3 | Three levels of information accuracy: high, intermediate, and low | Information compliance rate |
Tanaka et al. [39] | 15 | 60 | 2 | Sce. 1: LAI and without a penalty for late arrival Sce. 2: HAI and without a penalty for late arrival Sce. 3: LAI and a penalty for late arrival Sce. 4: HAI and a penalty for late arrival | Information compliance rate, choice rate of the shortest route |
Meneguzzer and Olivieri [40] | 30 | 50 | 3 | All participants only knew the TT of the chosen route and the initial free-flow time. | Average flow of three routes, average TT, frequency distribution of individual route switches, personal characteristics of participants |
Mak et al. [41] | 180 | 50 | 8 | The subjects were divided into 10 groups, five groups participated in condition RCC (provided route information) and five others in condition SCC (provided real-time segment information). | Flow distribution of routes and segments, route switching rate |
Zhao and Huang [42] | 18 | 30 | 2 | Provide feedback information to the subjects: the real and expected travel costs in the preceding round, their own road choice decisions in the preceding round, accumulative payoffs, and number of the current round. | Traffic flow distribution and TT cost of the two routes, proportion of choosing route M, perceived travel cost, and aspiration level of the subjects |
Wijayaratna et al. [31] | 12 | 20 | 3 | Exp. 1: no online information provided Exp. 2: online information provided at node C regarding the prevailing traffic conditions on link CB | Effect of online real-time information on the paradox |
Zhang et al. [43] | 25 | 135 | 3 | Subjects acted as commuters traveling, and the size of online travel communities which these subjects belong to increased from 0, 3 to 25 by arithmetic progression. | Effect of social interaction information from friends on the entire traffic system and individuals |
Klein and Ben-Elia [44] | 90 | 110 | 2 | First 10 rounds: participants were provided TT, arrival time, and payoff in the last round Next 50 rounds: participants were provided the previous experience information, expected TT, and their expected times of arrival Last 50 rounds: participants were provided the previous information and recommended route. | Collective response analysis, individual response analysis, response time analysis |
Ye et al. [45] | 268 | 26 | 3 | All participants were provided with the TT of all routes in the last round and the initial free-flow time. | Nonlinear effects of flow and cost differences on route switching, participant's path preference, time-varying sensitivity, and day-to-day learning behavior |
Qi et al. [46] | 64 | 80 | 2 | Same as Selten et al. [35]. | Difference between traffic flow and network equilibrium predictions, difference between the two information conditions, individual response modes |
|