Ant Colony Optimization Algorithm for understanding of trade-offs between safety and benefit: a case of Beijing taxi service system

Abstract

Creating and maintaining a high level of both safety and productivity is a primary objective for many industries, especially those endowed with pursuit of commercial profits which are tightly linked with higher exposure to occupational risks, such as commercial transportation service industry. As a typical case, Beijing taxi service system (BTSS) operates for trade-offs between safety and benefit but under a decentralized and loose control at the sharp-end level, which impels taxi drivers to tackle routine work in a highly cooperative manner, e.g., interacting, communicating and collaborating in local groups. Based on resemblances between the collective behavioral patterns of the driver groups and social insect systems, Ant Colony Optimization Algorithm (ACOA) is used to investigate mechanisms that coordinate the drivers’ individual efforts, e.g., recruit informational support when needed. The ACOA inference is validated subsequently with empirical evidence based on statistical analysis of the drivers’ attitude bias. Experimentation shows that the mathematical model of ACOA is successful in explaining how collective patterns of the drivers’ decisions are generated, as well as instantiates group-level resilience skills, with the drivers’ flexibly changing strategies towards the trade-offs in different scenarios where competition between safety and benefit escalates. The research findings suggest capacities of resilience and self-organization in the current BTSS, contributing to understandings of the safety-benefit trade-off mechanism that functionally integrates individuals at the sharp end of BTSS. The exploratory applications of ACOA to BTSS provide reference of improving risk and performance management in Beijing taxi service industry, as well as promote coherent research on human cognitive and behavioral properties in complex socio-technical systems.