# Key Simulation Results (Excel Spread Sheets)

## Synthetic Road Networks

##

Scenario: Small Synthetic Road Network
This scenario compares the performance of CARAVAN in terms of percentage reduction in travel time with three non-cooperative algorithms: Shortest path algorithm (where preferenceMultiplier = 0), preference-based algorithm (where costMultiplier = 0), and preference-based shortest path algorithm (where preferencemultiplier and costmultiplier are equal, and not equal to zero). The table below gives the aggregate travel time for the non-cooperative algorithms and for Social, Rational and Mixed welfares of CARAVAN.
Scenario: Large Synthetic Road Network
This scenario compares the performance of CARAVAN with the Shortest Path Algorithm in terms of percentage reduction in travel time for all the three welfare types with varying number of vehicles and junction count. The presented table shows the results for 1,3 and 5-junction scenarios.

## Small Real-road Network

##

Scenario: Evaluating Performance of CARAVAN in terms of Percentage Reduction in Travel Time
This scenario evaluates the performance of CARAVAN in terms of percentage reduction in travel time for the three welfare types of CARAVAN.
Scenario: Evaluating Performance of CARAVAN on basis of 5 parameters
This graph compares the performance of three welfare types in terms of Aggregate percentage reduction, Average percentage reduction, Best percentage reduction, Worst percentage reduction and the Standard Deviation in percentage reduction in travel time.

## Real-road grid Network

##

Scenario: Variation in Percentage of Non-equipped Vehicles
This scenario was simulated with a set of 40 to 90 vehicles. The percentage of non-equipped vehicles was varied from 25-100% for each of these sets and travel time was recorded. This travel time was compared to the travel time obtained when all the vehicles are equipped. The difference is expressed in terms of % Increase in travel time compared to all-equipped vehicles case.
Scenario: Variation in Percentage of Non-compliant Vehicles
This scenario was simulated with a set of 40 to 90 vehicles.The percentage of non-compliant vehicles was varied from 25-100% for each of these sets and travel time was recorded. This travel time was compared to the travel time obtained when all the vehicles are compliant. The difference is expressed in terms of % Increase in travel time compared to all-compliant vehicles case.
Scenario: Variation in Placement of Junctions
This scenario was tested with a set of 50 to 90 vehicles. This 4-junction scenario was tested for initial placement of junctions i.e. in the initial part of the road network (all 4 junctions at the beginning of the journey) and then making the junctions sparse (2 junctions each at the beginning and end of the journey). Results in terms of % Gain in travel time of setting 1 over setting 2 were obtained.
Scenario: Vehicles vs. Number of Junctions
This scenario is tested for sets of 50 to 90 vehicles for 2,4,6, and 8 junctions respectively. Each of the travel times obtained was compared with the Shortest Path Algorithm.
Scenario: Source-destination pair variation
Five scenarios with random sets of origin-destination pairs were simulated. The performance of CARAVAN (Social Welfare stategy) was compared with the Shortest Path algorithm in terms of percentage reduction in travel time for five simulation scenarios.

## Large Real-road Network

##

Scenario: Variation in Percentage of Non-equipped Vehicles
This scenario was simulated with a set of 60 to 80 vehicles. The percentage of non-equipped vehicles was varied from 25-100% for each of these sets and travel time was recorded. This travel time was compared to the travel time obtained when all the vehicles are equipped. The difference is expressed in terms of % Increase in travel time compared to all-equipped vehicles case.
Scenario: Number of Junctions
This scenario demonstrates the significance of repetitive cooperative decision-making on the performance of CARAVAN in terms of percentage reduction in travel time by varying the number of junctions (decision-making points). The percentage reduction in travel time as compared to the Shortest Path algorithm is recorded.
Scenario: Variation in Segment Capacity Threshold
The scenario demonstrates the effect of variations in segment threshold capacity on CARAVAN.By varying the segment capacity threshold between 7 and 14, the effect on percentage reduction in travel time over the Shortest Path Algorithm using CARAVAN is recorded.
Scenario: Congested Segments
This scenario demonstrates the ability of the algorithm to adapt its route allocation for a congested scenario in order to maximise the travel time savings for different values of p and c. For each of the scenarios with varying values of preference and cost multipliers, the percentage reduction in travel time as compared to the Shortest Path Algorithm is recorded.
Scenario: Variations in Preference Utility Weight Multipliers
In this scenario, the Preference Utility Weight multipliers were varied and their influence on the aggregate Preference Utility Index of the initial and final allocations were compared. The scenario analyses how the algorithm adapts to the varying significance of a preference criteria in Preference Utility Weight. The initial and final values of aggregate Preference Utility Index values is compared.

Back to Homepage