Ghassan Abu-Lebdeh

DEVELOPMENT OF TRAFFIC CONTROL AND QUEUE MANAGEMENT PROCEDURES FOR OVERSATURATED ARTERIALS

The formulation and solution of a new algorithm for queue management and coordination of traffic signals along oversaturated arterials are presented. Existing traffic-control and signal-coordination algorithms deal only with undersaturated steady-state traffic flow conditions. No practical algorithms are readily available for oversaturated flow conditions. The main idea of the procedure is to manage queue formation and dissipation on system links so that traffic flow is maximized by efficiently using all green time, preventing formation of de facto red, accounting for the non-steady-state conditions, and providing time-dependent control measures. The problem is formulated as a throughput maximization problem subject to state and control variables. The solution is then obtained using genetic algorithms. The results show that the control procedure can produce dynamic and responsive control so that traffic progression is attained and all undesirable conditions such as queue spill-back and de facto red are avoided.

GENETIC ALGORITHMS FOR TRAFFIC SIGNAL CONTROL AND QUEUE MANAGEMENT OF OVERSATURATED TWO-WAY ARTERIALS

The formulation and solution of a dynamic signal control and queue management algorithm for two-way oversaturated arterials are presented. The algorithm is structured to find optimal control and queue management in at least one direction: the primary direction. The other direction is optimized only subject to the fulfillment of the constraints on the primary direction. The main idea of the procedure is to manage queue formation and dissipation through proper design of signal control parameters such that queues are always contained within respective links and that spillbacks are prevented. The two-way control procedure presented accounts for all possible traffic flow regimes that may form as a result of queue management schemes. Genetic algorithms were used to solve the problem. The results show that although the algorithm provides optimal control and queue management in the primary direction, it was also able to manage queues in the secondary direction so that the occurrence of queue spillbacks is prevented. Progression was attainable only in the primary direction. However, the algorithm was able to effectively deal with evolving traffic queues in both directions. In all but one link, traffic was successfully contained in the respective links, thus preventing spillback.

DESIGN AND EVALUATION OF DYNAMIC TRAFFIC MANAGEMENT STRATEGIES FOR CONGESTED CONDITIONS

This paper presents a procedure for dynamic design and evaluation of traffic management strategies in oversaturated conditions. The method combines a dynamic control algorithm and a disutility function. The dynamic algorithm designs signal control parameters to manage formation and dissipation of queues on system links with explicit consideration of current and projected queue lengths and demands. The disutility function measures the relative performance of the dynamic control algorithm based on preset system performance goals. The user may statically select the management strategy, or alternatively the system may be instructed to set off different management schemes based on predefined performance thresholds. The problem was formulated as one of output maximization subject to state, control, and traffic management strategy choices. Solutions were obtained using genetic algorithms. Four traffic management plans were tested to show the capabilities of the new procedure. The results show that the procedure is able to generate suitable signal control schemes that are favorable to attaining the desired traffic management goals. The results showed that multiple, or hybrids of single measures of effectiveness may need to be examined in order to correctly assess system performance. The procedure has potential for real-time implementation in an intelligent transportation system setting.

Convergence variability and population sizing in micro-genetic algorithms

The issues of micro-genetic algorithm (micro-GA) convergence, population size, internal variability, and performance are examined. Procedures are developed to determine best population size when micro-GAs are used for optimization of real-world problems. Results show that the best population size cannot be determined before convergence has occurred and that convergence may occur without reaching an optimal solution. To determine the best population size, this article proposes the use of consistency of the final fitness values in addition to the fitness value at convergence. Internal variability must be considered in finding the best population size, and that best population size depends on the level of computational resources used. Smaller populations with generous computational resources performed as good as and sometimes better than larger populations. Midsized populations consistently exhibited lower variation in fitness value. For a given population size, the increase in fitness value may not be worth the added computation cost beyond a certain point. Three approaches are proposed to determine the best population size: (1) a random selection approach, (2) a detailed approach, and (3) a simplified approach. We used the detailed approach with our traffic problem and found that, with sufficient computational resources, the micro-GA performed best with population size around the square root of the string length. We tested this proposition on standard simple and deceptive problems and found it to hold true.

Signal Coordination and Arterial Capacity in Oversaturated Conditions

Models for estimation of the capacities of oversaturated arterials were developed. The input variables in these models are capacities of individual intersections, offsets, and vehicle queue lengths. Models for quantification of capacity loss due to blockage caused by downstream queues are also presented. The proposed models show that when arterial capacity is determined in oversaturated conditions, it is not sufficient to consider only the capacities of critical intersections; instead, the capacities of critical subsystems must be considered. A critical subsystem is any two intersections plus the link that joins them where traffic processing capability is the lowest. This traffic processing capability, or critical subsystem capacity, determines the arterial capacity. It is a function of the capacities of the respective intersections, the offset between them, and the queue length on the link joining them. It is shown that a critical subsystem is not unique in that it may change location over the course of the study period. To minimize capacity loss, it is shown that offsets must be an explicit function of queue lengths. The practical use of the models was demonstrated for an oversaturated two-intersection system. The results show that improper setting of offsets can lead to significant capacity loss. In extreme cases all capacity in a given cycle may be lost if the offsets are not set properly.

Modeling Traffic Output for Design of Dynamic Multi-Cycle Control in Congested Conditions

Accurate estimates of intermediate signal traffic output is necessary for designing meaningful multi-cycle intelligent or dynamic signal control in congested conditions. Models are presented that capture traffic output of intersections in congested interrupted flow conditions with explicit consideration of interactions, over space and time, between traffic streams at successive signals. Three unique operational characteristics along with their interactions were found to characterize distinct traffic flow regimes and determine system output. Models to predict the formation and intensity of these operational characteristics and the corresponding flow regimes were derived. Results demonstrate that an accurate representation of these regimes provides critical information for subsequent control steps.

Models for Right-Turn-on-Red and Their Effects on Intersection Delay

Models to estimate right-turn-on-red (RTOR) volumes at intersections with exclusive right-turn (RT) lanes are developed, and the effects of RTOR volumes on computed delay are assessed. The important variables in these models are the RT volume, followed by green-time-to-cycle (G/C) ratio, volume of conflicting traffic, and whether there is a protected phase for opposing left-turning vehicles. The estimated RTOR increased as the RTs increased. However, it decreased as G/C and the volume of conflicting traffic increased. Results show that not accounting for RTOR volumes can lead to a significant difference in delay estimates for RT lanes and, to a lesser extent, on the corresponding approaches. For RT lanes, in one-half of the cases the difference was greater than 5 percent, in more than one-quarter of the cases the difference was greater than 10 percent, and in at least one of eight cases the difference was greater than 20 percent. Differences for individual cases ranged between 0 and 130 percent, with an average of 12 percent. For approaches, the average delay difference was 4 percent, and for individual cases the difference ranged between −2 and 78 percent. As recommended by the Highway Capacity Manual, actual field counts of RTOR volume should be used whenever available. However, in the absence of such counts, the models developed here can be used and hence should be considered in capacity analysis procedures.

Computational issues in micro-genetic algorithms for traffic management

Emerging artificial intelligence techniques such as genetic algorithms (GAs) allow a more realistic representation and solution of difficult and combinatorial problems such as the dynamic traffic queue management problem. Computational experience in solving such complex large-scale problems by use of micro-GAs is described. In addition to providing evidence of the ability of micro-GAs to successfully identify optimal traffic management schemes, some micro-GA-associated computational issues that warrant attention are highlighted. Choosing a proper population size is a critical decision, and internal variability must be accounted for to assess the goodness of the optimization results properly. A simple rule for deciding the best population size for micro-GAs is proposed. Micro-GAs may converge to low-quality solutions, particularly with very small population sizes; convergence of micro-GAs by itself is not a sufficient indication of good performance. The size of the search space for some real-world systems can pose some difficulties to micro-GAs. Choosing between micro-GAs and regular GAs is a problem-dependent decision.

Configuring microgenetic algorithms for solving traffic control problems: the case of number of generations

Efficient and successful use of genetic algorithms (GAs) requires careful selection of several parameter values. One such critical parameter is the processing time (or, number of generations) that is sufficient to ensure suitable convergence. Todate there is only limited guidance on this subject, and in most cases detailed knowledge of the structure and properties of the problem is necessary for such guidance to be useable. For real world problems such knowledge may not be readily available. We describe an experimental approach to establish relationships between time to convergence and problem size of microgenetic algorithms (m-GAs). A discrete time dynamical traffic control problem with different sizes and levels of complexity was used as a test bed. The results showed that upon appropriately sizing the m-GA population, the m-GA can converge to a near-optimal solution in a number of generations equal to the string length. The results also demonstrate that with the selection of appropriate number of generations, it is possible to get most of the worth of the theoretically optimal solution but with only a fraction of the computation cost. The results showed that as the size of the optimization problem grew exponentially, the time requirements of m-GA grew only linearly thus making m-GAs especially suited for optimizing large scale and combinatorial problems for online optimization