Until recently, parking analysis procedures had remained traditional although analysis procedures for almost every other type of transportation infrastructure had evolved over the years. Empirical procedures of the Parking Generation Manual (PGM), developed by the Institute of Transportation Engineers (ITE), is still the most popular method for parking facility design. This method provides a starting ground for sizing a parking facility, but no performance measures can be computed.

Now we have SPARKS (Smart Parking Analysis Software) an innovative tool that is likely to change the way we conduct parking analyses. Based on scientific theory, SPARKS provides a variety of measures of effectiveness by which a facility can be evaluated. SPARKS uses the traffic operations method, which is based on quantitative analysis. This method had not been used extensively in the past because of, in part, the lack of an integrated tool.

Parking Operations

Two types of parking operations, on-street and off-street, have the following distinctly different operating characteristics and, therefore, should be modeled accordingly:

• Queuing and waiting delays often experienced in off-street facilities are negligible in on-street parking.
• Short-term parking is common in on-street facilities, but both long- and short-term parking are prevalent in off-street facilities.

Intelligent Transportation Systems (ITS) has provided incentives for advanced parking management systems (APMS). One such management technology—the Parking Guidance Systems (PGS)—minimizes searching delay in off-street facilities by guiding motorists directly to vacant spaces. There is a need for modeling APMS to assess benefits.

Modeling of on-street parking, off-street parking and parking management operations will allow assessment of the benefit-cost ratio that is required for a feasibility study. A proper evaluation of parking operations will minimize congestion, vehicular delay, fuel consumption and emissions, and reduce waste of resources.

How Smart is SPARKS?

Based on queuing models, SPARKS was modified to simulate parking operations. SPARKS represents the stochastic nature of vehicle arrival, random selection of spaces, delay in searching for a space, delay in departing from a space, queuing due to full occupancy, and vehicles leaving due to lack of a space. In all, there are four types of models in SPARKS:

• Off-street parking
• On-street parking
• On-street revenue generation
• On-street/off-street stochastic models for determination of the maximum number of vehicles parked.

A sample problem demonstrating an off-street parking application is presented below. Several other modeling techniques, including on-street and PGS operations, are documented in the SPARKS Users Manual.

Sample Problem

Consider that a shopping complex is being designed with a gross leasable area (GLA) of 53 627 square meters (175,940 square feet). A study detailing the number of parking spaces is to be submitted to the County for approval of the site plan.

Table 1: Input Data for Modeling Off-Street Parking Using SPARKS

Table 2: Measures of Effectiveness for Off-Street Parking

Parking Generation Manual Method. Consulting PGM, generation rates for weekday and weekend periods are found to be 3.23 and 3.97 spaces per 305 square meters (1,000 square feet) of GLA, respectively. Therefore, total parking demand will be:

53 627/305 (175,940/1,000) x 3.97 = 698 spaces.

Traffic Operations Method Using SPARKS. The sizing of the shopping center is next examined by the traffic operations method, which provides detailed performance measures. Three alternatives are studied:

1. Base case with parking supply estimated using PGM (2), with 698 spaces
2. A substantially lower estimate than PGM with 500 spaces
3. A slightly higher estimate, but still lower than PGM, with 575 spaces.

The collected input data and input definitions for SPARKS are shown in Table 1 on page 32. Measures of effectiveness and output definitions are shown in Table 2 on page 32.

As shown in Table 2, all MOEs for Alternative 1 are satisfactory. No waiting delay will be incurred and no vehicle will be waiting in the facility. The maximum number of vehicles in the facility will be 486, at 99 percent probability threshold, which is less than the parking supply of 698 spaces. The parking activity index of 1.10 shows a low level of activity.

Alternative 2 shows a negligible waiting delay will be incurred (0.00004 x 60 x 60 = 0.144 seconds) consisting of only 0.16 percent of arriving vehicles. The percent of arriving vehicles waiting more than the desired waiting period of 5 minutes is 0.01 with 1 vehicle waiting on an average. The maximum number of vehicles in the facility will be 486, which is less than the capacity of 500 with a maximum demand to supply ratio of 486/500 = 0.97. The PAI value of 2.14, higher than Alternative 1, indicates greater activity. Average occupancy increased to 0.87, which is a little high for this design. Overall, the facility operation is marginal.

Alternative 3 shows that no waiting delay will be incurred with no vehicle waiting. Also, no vehicle will wait greater than the desired period of 5 minutes. The maximum number of vehicles in the facility will be 486, with demand to supply ratio of 0.84. PAI value of 1.63, with an occupancy of 0.76, are satisfactory.

Alternatives 2 and 3 were considered to determine if the parking supply estimated by using PGM could be reduced without sacrificing the level of service. Results showed that Alternative 3 will be as effective as Alternative 1, with significantly lower supply. Therefore, it will be beneficial to choose 575 spaces as opposed to 698 spaces required by PGM. By constructing 123 fewer spaces, a substantial cost savings will be realized.

Conclusion. Unique concepts on parking operations for on-street and off-street facilities are presented. A four-step design procedure is suggested:

1. Develop initial parking supply based on land use type and size using PGM.
2. Use SPARKS to model traffic operations with the initial parking supply.
3. Adjust parking supply based on measures of effectiveness.
4. Simulate final design using SPARKS.

Software Information

Professor Dev Roy, formerly of the State University of New York at Utica/Rome, and I developed SPARKS, which is currently DOS-based. Dr. Roy is now considering several improvements to the software, such as:

• Enhancement to the Windows environment
• State-of-the-art input/output file management
• Database for parking inventory management.

SPARKS has been used in the U.S., Brazil and Taiwan. It is available through the McTrans Center in Florida and PcTrans in Kansas. Anyone within PB who may be interested in using this software can call me at 407-875-3337.